├── 1. Importing Data with Pandas.ipynb
├── 2. Data Preprocessing with Pandas.ipynb
├── 3. Feature Extraction.ipynb
├── 5. Handling Missing values.ipynb
├── 6. Data Joining.ipynb
├── 7. Pivot table.ipynb
├── 8. Categorical Variable Encoding.ipynb
├── ODI_cricket.xlsx
├── README.md
├── batsman_most_runs_ODI.csv
├── bmw.csv
├── bowler_most_wickets_ODI.csv
└── gre.csv


/1. Importing Data with Pandas.ipynb:
--------------------------------------------------------------------------------
  1 | {
  2 |   "cells": [
  3 |     {
  4 |       "cell_type": "markdown",
  5 |       "metadata": {
  6 |         "id": "view-in-github",
  7 |         "colab_type": "text"
  8 |       },
  9 |       "source": [
 10 |         "<a href=\"https://colab.research.google.com/github/SKawsar/data_preprocessing_for_ML_with_Python/blob/main/Lecture_01.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
 11 |       ]
 12 |     },
 13 |     {
 14 |       "cell_type": "markdown",
 15 |       "metadata": {
 16 |         "id": "Y66h391odn_x"
 17 |       },
 18 |       "source": [
 19 |         "# Lecture 01: Importing Data with Pandas\n",
 20 |         "\n",
 21 |         "Instructor: **Md Shahidullah Kawsar**\n",
 22 |         "<br>Data Scientist, IDARE, Houston, TX, USA\n",
 23 |         "\n",
 24 |         "#### Objectives:\n",
 25 |         "- Challenges of reading a CSV or Excel file\n",
 26 |         "- Choose columns by name before reading a csv file\n",
 27 |         "- Choose columns by number before reading a csv file\n",
 28 |         "- Reading only the first n number of rows\n",
 29 |         "\n",
 30 |         "#### References:\n",
 31 |         "[1] Data Source: https://stats.espncricinfo.com/ci/content/records/83548.html\n",
 32 |         "https://stats.espncricinfo.com/ci/content/records/283193.html\n",
 33 |         "<br>[2] Reading CSV file in pandas:\n",
 34 |         "https://pandas.pydata.org/docs/reference/api/pandas.read_csv.html\n",
 35 |         "<br>[3] Reading Excel file in pandas: https://pandas.pydata.org/docs/reference/api/pandas.read_excel.html"
 36 |       ]
 37 |     },
 38 |     {
 39 |       "cell_type": "markdown",
 40 |       "metadata": {
 41 |         "id": "EmvuNOxVdn_y"
 42 |       },
 43 |       "source": [
 44 |         "#### Import required libraries"
 45 |       ]
 46 |     },
 47 |     {
 48 |       "cell_type": "code",
 49 |       "execution_count": null,
 50 |       "metadata": {
 51 |         "id": "bdrNcRF0dn_y"
 52 |       },
 53 |       "outputs": [],
 54 |       "source": [
 55 |         "import pandas as pd"
 56 |       ]
 57 |     },
 58 |     {
 59 |       "cell_type": "code",
 60 |       "execution_count": null,
 61 |       "metadata": {
 62 |         "id": "824w1FcNdn_z",
 63 |         "outputId": "999b29a2-6067-46c5-f129-d53a2f0e962a"
 64 |       },
 65 |       "outputs": [
 66 |         {
 67 |           "data": {
 68 |             "text/html": [
 69 |               "<div>\n",
 70 |               "<style scoped>\n",
 71 |               "    .dataframe tbody tr th:only-of-type {\n",
 72 |               "        vertical-align: middle;\n",
 73 |               "    }\n",
 74 |               "\n",
 75 |               "    .dataframe tbody tr th {\n",
 76 |               "        vertical-align: top;\n",
 77 |               "    }\n",
 78 |               "\n",
 79 |               "    .dataframe thead th {\n",
 80 |               "        text-align: right;\n",
 81 |               "    }\n",
 82 |               "</style>\n",
 83 |               "<table border=\"1\" class=\"dataframe\">\n",
 84 |               "  <thead>\n",
 85 |               "    <tr style=\"text-align: right;\">\n",
 86 |               "      <th></th>\n",
 87 |               "      <th>Player</th>\n",
 88 |               "      <th>Span</th>\n",
 89 |               "      <th>Mat</th>\n",
 90 |               "      <th>Inns</th>\n",
 91 |               "      <th>NO</th>\n",
 92 |               "      <th>Runs</th>\n",
 93 |               "      <th>HS</th>\n",
 94 |               "      <th>Ave</th>\n",
 95 |               "      <th>BF</th>\n",
 96 |               "      <th>SR</th>\n",
 97 |               "      <th>100</th>\n",
 98 |               "      <th>50</th>\n",
 99 |               "      <th>0</th>\n",
100 |               "      <th>4s</th>\n",
101 |               "      <th>6s</th>\n",
102 |               "    </tr>\n",
103 |               "  </thead>\n",
104 |               "  <tbody>\n",
105 |               "    <tr>\n",
106 |               "      <th>0</th>\n",
107 |               "      <td>SR Tendulkar (INDIA)</td>\n",
108 |               "      <td>1989-2012</td>\n",
109 |               "      <td>463</td>\n",
110 |               "      <td>452</td>\n",
111 |               "      <td>41</td>\n",
112 |               "      <td>18426</td>\n",
113 |               "      <td>200*</td>\n",
114 |               "      <td>44.83</td>\n",
115 |               "      <td>21368</td>\n",
116 |               "      <td>86.23</td>\n",
117 |               "      <td>49</td>\n",
118 |               "      <td>96</td>\n",
119 |               "      <td>20</td>\n",
120 |               "      <td>2016</td>\n",
121 |               "      <td>195</td>\n",
122 |               "    </tr>\n",
123 |               "    <tr>\n",
124 |               "      <th>1</th>\n",
125 |               "      <td>KC Sangakkara (Asia/ICC/SL)</td>\n",
126 |               "      <td>2000-2015</td>\n",
127 |               "      <td>404</td>\n",
128 |               "      <td>380</td>\n",
129 |               "      <td>41</td>\n",
130 |               "      <td>14234</td>\n",
131 |               "      <td>169</td>\n",
132 |               "      <td>41.98</td>\n",
133 |               "      <td>18048</td>\n",
134 |               "      <td>78.86</td>\n",
135 |               "      <td>25</td>\n",
136 |               "      <td>93</td>\n",
137 |               "      <td>15</td>\n",
138 |               "      <td>1385</td>\n",
139 |               "      <td>88</td>\n",
140 |               "    </tr>\n",
141 |               "    <tr>\n",
142 |               "      <th>2</th>\n",
143 |               "      <td>RT Ponting (AUS/ICC)</td>\n",
144 |               "      <td>1995-2012</td>\n",
145 |               "      <td>375</td>\n",
146 |               "      <td>365</td>\n",
147 |               "      <td>39</td>\n",
148 |               "      <td>13704</td>\n",
149 |               "      <td>164</td>\n",
150 |               "      <td>42.03</td>\n",
151 |               "      <td>17046</td>\n",
152 |               "      <td>80.39</td>\n",
153 |               "      <td>30</td>\n",
154 |               "      <td>82</td>\n",
155 |               "      <td>20</td>\n",
156 |               "      <td>1231</td>\n",
157 |               "      <td>162</td>\n",
158 |               "    </tr>\n",
159 |               "    <tr>\n",
160 |               "      <th>3</th>\n",
161 |               "      <td>ST Jayasuriya (Asia/SL)</td>\n",
162 |               "      <td>1989-2011</td>\n",
163 |               "      <td>445</td>\n",
164 |               "      <td>433</td>\n",
165 |               "      <td>18</td>\n",
166 |               "      <td>13430</td>\n",
167 |               "      <td>189</td>\n",
168 |               "      <td>32.36</td>\n",
169 |               "      <td>14725</td>\n",
170 |               "      <td>91.20</td>\n",
171 |               "      <td>28</td>\n",
172 |               "      <td>68</td>\n",
173 |               "      <td>34</td>\n",
174 |               "      <td>1500</td>\n",
175 |               "      <td>270</td>\n",
176 |               "    </tr>\n",
177 |               "    <tr>\n",
178 |               "      <th>4</th>\n",
179 |               "      <td>DPMD Jayawardene (Asia/SL)</td>\n",
180 |               "      <td>1998-2015</td>\n",
181 |               "      <td>448</td>\n",
182 |               "      <td>418</td>\n",
183 |               "      <td>39</td>\n",
184 |               "      <td>12650</td>\n",
185 |               "      <td>144</td>\n",
186 |               "      <td>33.37</td>\n",
187 |               "      <td>16020</td>\n",
188 |               "      <td>78.96</td>\n",
189 |               "      <td>19</td>\n",
190 |               "      <td>77</td>\n",
191 |               "      <td>28</td>\n",
192 |               "      <td>1119</td>\n",
193 |               "      <td>76</td>\n",
194 |               "    </tr>\n",
195 |               "  </tbody>\n",
196 |               "</table>\n",
197 |               "</div>"
198 |             ],
199 |             "text/plain": [
200 |               "                        Player       Span  Mat  Inns  NO   Runs    HS    Ave  \\\n",
201 |               "0         SR Tendulkar (INDIA)  1989-2012  463   452  41  18426  200*  44.83   \n",
202 |               "1  KC Sangakkara (Asia/ICC/SL)  2000-2015  404   380  41  14234   169  41.98   \n",
203 |               "2         RT Ponting (AUS/ICC)  1995-2012  375   365  39  13704   164  42.03   \n",
204 |               "3      ST Jayasuriya (Asia/SL)  1989-2011  445   433  18  13430   189  32.36   \n",
205 |               "4   DPMD Jayawardene (Asia/SL)  1998-2015  448   418  39  12650   144  33.37   \n",
206 |               "\n",
207 |               "      BF     SR  100  50   0    4s   6s  \n",
208 |               "0  21368  86.23   49  96  20  2016  195  \n",
209 |               "1  18048  78.86   25  93  15  1385   88  \n",
210 |               "2  17046  80.39   30  82  20  1231  162  \n",
211 |               "3  14725  91.20   28  68  34  1500  270  \n",
212 |               "4  16020  78.96   19  77  28  1119   76  "
213 |             ]
214 |           },
215 |           "metadata": {},
216 |           "output_type": "display_data"
217 |         },
218 |         {
219 |           "name": "stdout",
220 |           "output_type": "stream",
221 |           "text": [
222 |             "(92, 15)\n"
223 |           ]
224 |         }
225 |       ],
226 |       "source": [
227 |         "# method 1\n",
228 |         "# reading a csv file \n",
229 |         "df = pd.read_csv(\"batsman_most_runs_ODI.csv\")\n",
230 |         "\n",
231 |         "display(df.head())\n",
232 |         "# print(df.tail())\n",
233 |         "print(df.shape)"
234 |       ]
235 |     },
236 |     {
237 |       "cell_type": "code",
238 |       "execution_count": null,
239 |       "metadata": {
240 |         "id": "3kezCVUodn_0",
241 |         "outputId": "bccb0c9e-e12c-4136-8d8b-c483f1b9c789"
242 |       },
243 |       "outputs": [
244 |         {
245 |           "name": "stdout",
246 |           "output_type": "stream",
247 |           "text": [
248 |             "number of rows =  92\n",
249 |             "number of columns =  15\n"
250 |           ]
251 |         }
252 |       ],
253 |       "source": [
254 |         "# number of rows\n",
255 |         "print(\"number of rows = \", df.shape[0])\n",
256 |         "\n",
257 |         "# number of columns\n",
258 |         "print(\"number of columns = \", df.shape[1])"
259 |       ]
260 |     },
261 |     {
262 |       "cell_type": "code",
263 |       "execution_count": null,
264 |       "metadata": {
265 |         "id": "l31EuE1Hdn_0",
266 |         "outputId": "b1ea11b1-fb12-40f6-8ce7-4d9e21a406fa"
267 |       },
268 |       "outputs": [
269 |         {
270 |           "name": "stdout",
271 |           "output_type": "stream",
272 |           "text": [
273 |             "<class 'pandas.core.frame.DataFrame'>\n",
274 |             "RangeIndex: 92 entries, 0 to 91\n",
275 |             "Data columns (total 15 columns):\n",
276 |             " #   Column  Non-Null Count  Dtype  \n",
277 |             "---  ------  --------------  -----  \n",
278 |             " 0   Player  92 non-null     object \n",
279 |             " 1   Span    92 non-null     object \n",
280 |             " 2   Mat     92 non-null     int64  \n",
281 |             " 3   Inns    92 non-null     int64  \n",
282 |             " 4   NO      92 non-null     int64  \n",
283 |             " 5   Runs    92 non-null     int64  \n",
284 |             " 6   HS      92 non-null     object \n",
285 |             " 7   Ave     92 non-null     float64\n",
286 |             " 8   BF      92 non-null     int64  \n",
287 |             " 9   SR      92 non-null     float64\n",
288 |             " 10  100     92 non-null     int64  \n",
289 |             " 11  50      92 non-null     int64  \n",
290 |             " 12  0       92 non-null     int64  \n",
291 |             " 13  4s      92 non-null     object \n",
292 |             " 14  6s      92 non-null     object \n",
293 |             "dtypes: float64(2), int64(8), object(5)\n",
294 |             "memory usage: 10.9+ KB\n",
295 |             "None\n"
296 |           ]
297 |         }
298 |       ],
299 |       "source": [
300 |         "# checking for missing values and data types of each column\n",
301 |         "print(df.info())"
302 |       ]
303 |     },
304 |     {
305 |       "cell_type": "code",
306 |       "execution_count": null,
307 |       "metadata": {
308 |         "id": "V8hS9RPfdn_0",
309 |         "outputId": "26c9a813-0bde-4bbf-88c7-90590af27291"
310 |       },
311 |       "outputs": [
312 |         {
313 |           "data": {
314 |             "text/html": [
315 |               "<div>\n",
316 |               "<style scoped>\n",
317 |               "    .dataframe tbody tr th:only-of-type {\n",
318 |               "        vertical-align: middle;\n",
319 |               "    }\n",
320 |               "\n",
321 |               "    .dataframe tbody tr th {\n",
322 |               "        vertical-align: top;\n",
323 |               "    }\n",
324 |               "\n",
325 |               "    .dataframe thead th {\n",
326 |               "        text-align: right;\n",
327 |               "    }\n",
328 |               "</style>\n",
329 |               "<table border=\"1\" class=\"dataframe\">\n",
330 |               "  <thead>\n",
331 |               "    <tr style=\"text-align: right;\">\n",
332 |               "      <th></th>\n",
333 |               "      <th>Mat</th>\n",
334 |               "      <th>Inns</th>\n",
335 |               "      <th>NO</th>\n",
336 |               "      <th>Runs</th>\n",
337 |               "      <th>Ave</th>\n",
338 |               "      <th>BF</th>\n",
339 |               "      <th>SR</th>\n",
340 |               "      <th>100</th>\n",
341 |               "      <th>50</th>\n",
342 |               "      <th>0</th>\n",
343 |               "    </tr>\n",
344 |               "  </thead>\n",
345 |               "  <tbody>\n",
346 |               "    <tr>\n",
347 |               "      <th>count</th>\n",
348 |               "      <td>92.000000</td>\n",
349 |               "      <td>92.000000</td>\n",
350 |               "      <td>92.000000</td>\n",
351 |               "      <td>92.000000</td>\n",
352 |               "      <td>92.000000</td>\n",
353 |               "      <td>92.000000</td>\n",
354 |               "      <td>92.000000</td>\n",
355 |               "      <td>92.000000</td>\n",
356 |               "      <td>92.000000</td>\n",
357 |               "      <td>92.000000</td>\n",
358 |               "    </tr>\n",
359 |               "    <tr>\n",
360 |               "      <th>mean</th>\n",
361 |               "      <td>241.021739</td>\n",
362 |               "      <td>226.891304</td>\n",
363 |               "      <td>29.217391</td>\n",
364 |               "      <td>7617.880435</td>\n",
365 |               "      <td>39.145761</td>\n",
366 |               "      <td>9560.043478</td>\n",
367 |               "      <td>80.414239</td>\n",
368 |               "      <td>12.217391</td>\n",
369 |               "      <td>46.043478</td>\n",
370 |               "      <td>12.847826</td>\n",
371 |               "    </tr>\n",
372 |               "    <tr>\n",
373 |               "      <th>std</th>\n",
374 |               "      <td>73.548964</td>\n",
375 |               "      <td>69.194813</td>\n",
376 |               "      <td>15.413585</td>\n",
377 |               "      <td>2490.710229</td>\n",
378 |               "      <td>6.264166</td>\n",
379 |               "      <td>3112.855969</td>\n",
380 |               "      <td>9.870732</td>\n",
381 |               "      <td>8.580358</td>\n",
382 |               "      <td>15.915928</td>\n",
383 |               "      <td>6.101585</td>\n",
384 |               "    </tr>\n",
385 |               "    <tr>\n",
386 |               "      <th>min</th>\n",
387 |               "      <td>128.000000</td>\n",
388 |               "      <td>126.000000</td>\n",
389 |               "      <td>3.000000</td>\n",
390 |               "      <td>5047.000000</td>\n",
391 |               "      <td>23.570000</td>\n",
392 |               "      <td>5504.000000</td>\n",
393 |               "      <td>60.570000</td>\n",
394 |               "      <td>0.000000</td>\n",
395 |               "      <td>23.000000</td>\n",
396 |               "      <td>2.000000</td>\n",
397 |               "    </tr>\n",
398 |               "    <tr>\n",
399 |               "      <th>25%</th>\n",
400 |               "      <td>189.500000</td>\n",
401 |               "      <td>179.000000</td>\n",
402 |               "      <td>16.000000</td>\n",
403 |               "      <td>5760.000000</td>\n",
404 |               "      <td>35.017500</td>\n",
405 |               "      <td>7422.250000</td>\n",
406 |               "      <td>74.125000</td>\n",
407 |               "      <td>6.000000</td>\n",
408 |               "      <td>35.000000</td>\n",
409 |               "      <td>8.000000</td>\n",
410 |               "    </tr>\n",
411 |               "    <tr>\n",
412 |               "      <th>50%</th>\n",
413 |               "      <td>229.000000</td>\n",
414 |               "      <td>217.500000</td>\n",
415 |               "      <td>29.500000</td>\n",
416 |               "      <td>6855.000000</td>\n",
417 |               "      <td>38.730000</td>\n",
418 |               "      <td>9144.500000</td>\n",
419 |               "      <td>78.960000</td>\n",
420 |               "      <td>10.500000</td>\n",
421 |               "      <td>41.500000</td>\n",
422 |               "      <td>13.000000</td>\n",
423 |               "    </tr>\n",
424 |               "    <tr>\n",
425 |               "      <th>75%</th>\n",
426 |               "      <td>274.750000</td>\n",
427 |               "      <td>256.500000</td>\n",
428 |               "      <td>40.000000</td>\n",
429 |               "      <td>8661.250000</td>\n",
430 |               "      <td>42.340000</td>\n",
431 |               "      <td>10957.000000</td>\n",
432 |               "      <td>87.167500</td>\n",
433 |               "      <td>16.250000</td>\n",
434 |               "      <td>52.250000</td>\n",
435 |               "      <td>16.000000</td>\n",
436 |               "    </tr>\n",
437 |               "    <tr>\n",
438 |               "      <th>max</th>\n",
439 |               "      <td>463.000000</td>\n",
440 |               "      <td>452.000000</td>\n",
441 |               "      <td>84.000000</td>\n",
442 |               "      <td>18426.000000</td>\n",
443 |               "      <td>58.070000</td>\n",
444 |               "      <td>21368.000000</td>\n",
445 |               "      <td>117.000000</td>\n",
446 |               "      <td>49.000000</td>\n",
447 |               "      <td>96.000000</td>\n",
448 |               "      <td>34.000000</td>\n",
449 |               "    </tr>\n",
450 |               "  </tbody>\n",
451 |               "</table>\n",
452 |               "</div>"
453 |             ],
454 |             "text/plain": [
455 |               "              Mat        Inns         NO          Runs        Ave  \\\n",
456 |               "count   92.000000   92.000000  92.000000     92.000000  92.000000   \n",
457 |               "mean   241.021739  226.891304  29.217391   7617.880435  39.145761   \n",
458 |               "std     73.548964   69.194813  15.413585   2490.710229   6.264166   \n",
459 |               "min    128.000000  126.000000   3.000000   5047.000000  23.570000   \n",
460 |               "25%    189.500000  179.000000  16.000000   5760.000000  35.017500   \n",
461 |               "50%    229.000000  217.500000  29.500000   6855.000000  38.730000   \n",
462 |               "75%    274.750000  256.500000  40.000000   8661.250000  42.340000   \n",
463 |               "max    463.000000  452.000000  84.000000  18426.000000  58.070000   \n",
464 |               "\n",
465 |               "                 BF          SR        100         50          0  \n",
466 |               "count     92.000000   92.000000  92.000000  92.000000  92.000000  \n",
467 |               "mean    9560.043478   80.414239  12.217391  46.043478  12.847826  \n",
468 |               "std     3112.855969    9.870732   8.580358  15.915928   6.101585  \n",
469 |               "min     5504.000000   60.570000   0.000000  23.000000   2.000000  \n",
470 |               "25%     7422.250000   74.125000   6.000000  35.000000   8.000000  \n",
471 |               "50%     9144.500000   78.960000  10.500000  41.500000  13.000000  \n",
472 |               "75%    10957.000000   87.167500  16.250000  52.250000  16.000000  \n",
473 |               "max    21368.000000  117.000000  49.000000  96.000000  34.000000  "
474 |             ]
475 |           },
476 |           "metadata": {},
477 |           "output_type": "display_data"
478 |         }
479 |       ],
480 |       "source": [
481 |         "# checking data statistics\n",
482 |         "display(df.describe())"
483 |       ]
484 |     },
485 |     {
486 |       "cell_type": "code",
487 |       "execution_count": null,
488 |       "metadata": {
489 |         "id": "ioXEtL3ldn_0",
490 |         "outputId": "5662e3b0-b91e-40a6-f0da-d0c38970c9cb"
491 |       },
492 |       "outputs": [
493 |         {
494 |           "name": "stdout",
495 |           "output_type": "stream",
496 |           "text": [
497 |             "Index(['Player', 'Span', 'Mat', 'Inns', 'NO', 'Runs', 'HS', 'Ave', 'BF', 'SR',\n",
498 |             "       '100', '50', '0', '4s', '6s'],\n",
499 |             "      dtype='object')\n"
500 |           ]
501 |         }
502 |       ],
503 |       "source": [
504 |         "# column names\n",
505 |         "print(df.columns)"
506 |       ]
507 |     },
508 |     {
509 |       "cell_type": "code",
510 |       "execution_count": null,
511 |       "metadata": {
512 |         "id": "Y8k5_uuIdn_3"
513 |       },
514 |       "outputs": [],
515 |       "source": [
516 |         ""
517 |       ]
518 |     }
519 |   ],
520 |   "metadata": {
521 |     "colab": {
522 |       "name": "Lecture_01.ipynb",
523 |       "provenance": [],
524 |       "include_colab_link": true
525 |     },
526 |     "kernelspec": {
527 |       "display_name": "Python 3",
528 |       "language": "python",
529 |       "name": "python3"
530 |     },
531 |     "language_info": {
532 |       "codemirror_mode": {
533 |         "name": "ipython",
534 |         "version": 3
535 |       },
536 |       "file_extension": ".py",
537 |       "mimetype": "text/x-python",
538 |       "name": "python",
539 |       "nbconvert_exporter": "python",
540 |       "pygments_lexer": "ipython3",
541 |       "version": "3.7.7"
542 |     }
543 |   },
544 |   "nbformat": 4,
545 |   "nbformat_minor": 0
546 | }
547 | 


--------------------------------------------------------------------------------
/2. Data Preprocessing with Pandas.ipynb:
--------------------------------------------------------------------------------
   1 | {
   2 |   "cells": [
   3 |     {
   4 |       "cell_type": "markdown",
   5 |       "metadata": {
   6 |         "id": "view-in-github",
   7 |         "colab_type": "text"
   8 |       },
   9 |       "source": [
  10 |         "<a href=\"https://colab.research.google.com/github/SKawsar/data_preprocessing_for_ML_with_Python/blob/main/Lecture_02.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
  11 |       ]
  12 |     },
  13 |     {
  14 |       "cell_type": "markdown",
  15 |       "metadata": {
  16 |         "id": "Y66h391odn_x"
  17 |       },
  18 |       "source": [
  19 |         "# Lecture 02: Data Preprocessing with Pandas (Part 1)\n",
  20 |         "\n",
  21 |         "Instructor: **Md Shahidullah Kawsar**\n",
  22 |         "<br>Data Scientist, IDARE, Houston, TX, USA\n",
  23 |         "\n",
  24 |         "#### Objectives:\n",
  25 |         "- Challenges of reading a CSV or Excel file\n",
  26 |         "- Choose columns by name before reading a csv file\n",
  27 |         "- Choose columns by number before reading a csv file\n",
  28 |         "- Reading only the first n number of rows\n",
  29 |         "\n",
  30 |         "#### References:\n",
  31 |         "[1] Data Source: https://stats.espncricinfo.com/ci/content/records/83548.html\n",
  32 |         "https://stats.espncricinfo.com/ci/content/records/283193.html\n",
  33 |         "<br>[2] Reading CSV file in pandas:\n",
  34 |         "https://pandas.pydata.org/docs/reference/api/pandas.read_csv.html\n",
  35 |         "<br>[3] Reading Excel file in pandas: https://pandas.pydata.org/docs/reference/api/pandas.read_excel.html"
  36 |       ]
  37 |     },
  38 |     {
  39 |       "cell_type": "markdown",
  40 |       "metadata": {
  41 |         "id": "EmvuNOxVdn_y"
  42 |       },
  43 |       "source": [
  44 |         "#### Import required libraries"
  45 |       ]
  46 |     },
  47 |     {
  48 |       "cell_type": "code",
  49 |       "execution_count": null,
  50 |       "metadata": {
  51 |         "id": "bdrNcRF0dn_y"
  52 |       },
  53 |       "outputs": [],
  54 |       "source": [
  55 |         "import pandas as pd"
  56 |       ]
  57 |     },
  58 |     {
  59 |       "cell_type": "markdown",
  60 |       "metadata": {
  61 |         "id": "rMU5Wi_idn_1"
  62 |       },
  63 |       "source": [
  64 |         "#### choose columns by name to read a csv file"
  65 |       ]
  66 |     },
  67 |     {
  68 |       "cell_type": "code",
  69 |       "execution_count": null,
  70 |       "metadata": {
  71 |         "id": "0aA0-7Fzdn_1",
  72 |         "outputId": "0b873b89-2bcd-40a6-87ec-7f3b8fd2a35d"
  73 |       },
  74 |       "outputs": [
  75 |         {
  76 |           "data": {
  77 |             "text/html": [
  78 |               "<div>\n",
  79 |               "<style scoped>\n",
  80 |               "    .dataframe tbody tr th:only-of-type {\n",
  81 |               "        vertical-align: middle;\n",
  82 |               "    }\n",
  83 |               "\n",
  84 |               "    .dataframe tbody tr th {\n",
  85 |               "        vertical-align: top;\n",
  86 |               "    }\n",
  87 |               "\n",
  88 |               "    .dataframe thead th {\n",
  89 |               "        text-align: right;\n",
  90 |               "    }\n",
  91 |               "</style>\n",
  92 |               "<table border=\"1\" class=\"dataframe\">\n",
  93 |               "  <thead>\n",
  94 |               "    <tr style=\"text-align: right;\">\n",
  95 |               "      <th></th>\n",
  96 |               "      <th>Player</th>\n",
  97 |               "      <th>Mat</th>\n",
  98 |               "      <th>Runs</th>\n",
  99 |               "      <th>100</th>\n",
 100 |               "    </tr>\n",
 101 |               "  </thead>\n",
 102 |               "  <tbody>\n",
 103 |               "    <tr>\n",
 104 |               "      <th>0</th>\n",
 105 |               "      <td>SR Tendulkar (INDIA)</td>\n",
 106 |               "      <td>463</td>\n",
 107 |               "      <td>18426</td>\n",
 108 |               "      <td>49</td>\n",
 109 |               "    </tr>\n",
 110 |               "    <tr>\n",
 111 |               "      <th>1</th>\n",
 112 |               "      <td>KC Sangakkara (Asia/ICC/SL)</td>\n",
 113 |               "      <td>404</td>\n",
 114 |               "      <td>14234</td>\n",
 115 |               "      <td>25</td>\n",
 116 |               "    </tr>\n",
 117 |               "    <tr>\n",
 118 |               "      <th>2</th>\n",
 119 |               "      <td>RT Ponting (AUS/ICC)</td>\n",
 120 |               "      <td>375</td>\n",
 121 |               "      <td>13704</td>\n",
 122 |               "      <td>30</td>\n",
 123 |               "    </tr>\n",
 124 |               "    <tr>\n",
 125 |               "      <th>3</th>\n",
 126 |               "      <td>ST Jayasuriya (Asia/SL)</td>\n",
 127 |               "      <td>445</td>\n",
 128 |               "      <td>13430</td>\n",
 129 |               "      <td>28</td>\n",
 130 |               "    </tr>\n",
 131 |               "    <tr>\n",
 132 |               "      <th>4</th>\n",
 133 |               "      <td>DPMD Jayawardene (Asia/SL)</td>\n",
 134 |               "      <td>448</td>\n",
 135 |               "      <td>12650</td>\n",
 136 |               "      <td>19</td>\n",
 137 |               "    </tr>\n",
 138 |               "    <tr>\n",
 139 |               "      <th>5</th>\n",
 140 |               "      <td>V Kohli (INDIA)</td>\n",
 141 |               "      <td>260</td>\n",
 142 |               "      <td>12311</td>\n",
 143 |               "      <td>43</td>\n",
 144 |               "    </tr>\n",
 145 |               "    <tr>\n",
 146 |               "      <th>6</th>\n",
 147 |               "      <td>Inzamam-ul-Haq (Asia/PAK)</td>\n",
 148 |               "      <td>378</td>\n",
 149 |               "      <td>11739</td>\n",
 150 |               "      <td>10</td>\n",
 151 |               "    </tr>\n",
 152 |               "    <tr>\n",
 153 |               "      <th>7</th>\n",
 154 |               "      <td>JH Kallis (Afr/ICC/SA)</td>\n",
 155 |               "      <td>328</td>\n",
 156 |               "      <td>11579</td>\n",
 157 |               "      <td>17</td>\n",
 158 |               "    </tr>\n",
 159 |               "    <tr>\n",
 160 |               "      <th>8</th>\n",
 161 |               "      <td>SC Ganguly (Asia/INDIA)</td>\n",
 162 |               "      <td>311</td>\n",
 163 |               "      <td>11363</td>\n",
 164 |               "      <td>22</td>\n",
 165 |               "    </tr>\n",
 166 |               "    <tr>\n",
 167 |               "      <th>9</th>\n",
 168 |               "      <td>R Dravid (Asia/ICC/INDIA)</td>\n",
 169 |               "      <td>344</td>\n",
 170 |               "      <td>10889</td>\n",
 171 |               "      <td>12</td>\n",
 172 |               "    </tr>\n",
 173 |               "  </tbody>\n",
 174 |               "</table>\n",
 175 |               "</div>"
 176 |             ],
 177 |             "text/plain": [
 178 |               "                        Player  Mat   Runs  100\n",
 179 |               "0         SR Tendulkar (INDIA)  463  18426   49\n",
 180 |               "1  KC Sangakkara (Asia/ICC/SL)  404  14234   25\n",
 181 |               "2         RT Ponting (AUS/ICC)  375  13704   30\n",
 182 |               "3      ST Jayasuriya (Asia/SL)  445  13430   28\n",
 183 |               "4   DPMD Jayawardene (Asia/SL)  448  12650   19\n",
 184 |               "5              V Kohli (INDIA)  260  12311   43\n",
 185 |               "6    Inzamam-ul-Haq (Asia/PAK)  378  11739   10\n",
 186 |               "7       JH Kallis (Afr/ICC/SA)  328  11579   17\n",
 187 |               "8      SC Ganguly (Asia/INDIA)  311  11363   22\n",
 188 |               "9    R Dravid (Asia/ICC/INDIA)  344  10889   12"
 189 |             ]
 190 |           },
 191 |           "metadata": {},
 192 |           "output_type": "display_data"
 193 |         }
 194 |       ],
 195 |       "source": [
 196 |         "col_names = ['Player', 'Mat', 'Runs', '100']\n",
 197 |         "df_usecols = pd.read_csv(\"batsman_most_runs_ODI.csv\", usecols=col_names)\n",
 198 |         "\n",
 199 |         "display(df_usecols.head(10))"
 200 |       ]
 201 |     },
 202 |     {
 203 |       "cell_type": "markdown",
 204 |       "metadata": {
 205 |         "id": "Fs9HUH6Hdn_1"
 206 |       },
 207 |       "source": [
 208 |         "#### Choose columns by number to read a csv file"
 209 |       ]
 210 |     },
 211 |     {
 212 |       "cell_type": "code",
 213 |       "execution_count": null,
 214 |       "metadata": {
 215 |         "id": "u_86zNcHdn_2",
 216 |         "outputId": "eed640f3-5fe2-48a9-99e7-3311d9a0488e"
 217 |       },
 218 |       "outputs": [
 219 |         {
 220 |           "data": {
 221 |             "text/html": [
 222 |               "<div>\n",
 223 |               "<style scoped>\n",
 224 |               "    .dataframe tbody tr th:only-of-type {\n",
 225 |               "        vertical-align: middle;\n",
 226 |               "    }\n",
 227 |               "\n",
 228 |               "    .dataframe tbody tr th {\n",
 229 |               "        vertical-align: top;\n",
 230 |               "    }\n",
 231 |               "\n",
 232 |               "    .dataframe thead th {\n",
 233 |               "        text-align: right;\n",
 234 |               "    }\n",
 235 |               "</style>\n",
 236 |               "<table border=\"1\" class=\"dataframe\">\n",
 237 |               "  <thead>\n",
 238 |               "    <tr style=\"text-align: right;\">\n",
 239 |               "      <th></th>\n",
 240 |               "      <th>Player</th>\n",
 241 |               "      <th>Mat</th>\n",
 242 |               "      <th>Runs</th>\n",
 243 |               "      <th>100</th>\n",
 244 |               "    </tr>\n",
 245 |               "  </thead>\n",
 246 |               "  <tbody>\n",
 247 |               "    <tr>\n",
 248 |               "      <th>0</th>\n",
 249 |               "      <td>SR Tendulkar (INDIA)</td>\n",
 250 |               "      <td>463</td>\n",
 251 |               "      <td>18426</td>\n",
 252 |               "      <td>49</td>\n",
 253 |               "    </tr>\n",
 254 |               "    <tr>\n",
 255 |               "      <th>1</th>\n",
 256 |               "      <td>KC Sangakkara (Asia/ICC/SL)</td>\n",
 257 |               "      <td>404</td>\n",
 258 |               "      <td>14234</td>\n",
 259 |               "      <td>25</td>\n",
 260 |               "    </tr>\n",
 261 |               "    <tr>\n",
 262 |               "      <th>2</th>\n",
 263 |               "      <td>RT Ponting (AUS/ICC)</td>\n",
 264 |               "      <td>375</td>\n",
 265 |               "      <td>13704</td>\n",
 266 |               "      <td>30</td>\n",
 267 |               "    </tr>\n",
 268 |               "    <tr>\n",
 269 |               "      <th>3</th>\n",
 270 |               "      <td>ST Jayasuriya (Asia/SL)</td>\n",
 271 |               "      <td>445</td>\n",
 272 |               "      <td>13430</td>\n",
 273 |               "      <td>28</td>\n",
 274 |               "    </tr>\n",
 275 |               "    <tr>\n",
 276 |               "      <th>4</th>\n",
 277 |               "      <td>DPMD Jayawardene (Asia/SL)</td>\n",
 278 |               "      <td>448</td>\n",
 279 |               "      <td>12650</td>\n",
 280 |               "      <td>19</td>\n",
 281 |               "    </tr>\n",
 282 |               "    <tr>\n",
 283 |               "      <th>5</th>\n",
 284 |               "      <td>V Kohli (INDIA)</td>\n",
 285 |               "      <td>260</td>\n",
 286 |               "      <td>12311</td>\n",
 287 |               "      <td>43</td>\n",
 288 |               "    </tr>\n",
 289 |               "    <tr>\n",
 290 |               "      <th>6</th>\n",
 291 |               "      <td>Inzamam-ul-Haq (Asia/PAK)</td>\n",
 292 |               "      <td>378</td>\n",
 293 |               "      <td>11739</td>\n",
 294 |               "      <td>10</td>\n",
 295 |               "    </tr>\n",
 296 |               "    <tr>\n",
 297 |               "      <th>7</th>\n",
 298 |               "      <td>JH Kallis (Afr/ICC/SA)</td>\n",
 299 |               "      <td>328</td>\n",
 300 |               "      <td>11579</td>\n",
 301 |               "      <td>17</td>\n",
 302 |               "    </tr>\n",
 303 |               "    <tr>\n",
 304 |               "      <th>8</th>\n",
 305 |               "      <td>SC Ganguly (Asia/INDIA)</td>\n",
 306 |               "      <td>311</td>\n",
 307 |               "      <td>11363</td>\n",
 308 |               "      <td>22</td>\n",
 309 |               "    </tr>\n",
 310 |               "    <tr>\n",
 311 |               "      <th>9</th>\n",
 312 |               "      <td>R Dravid (Asia/ICC/INDIA)</td>\n",
 313 |               "      <td>344</td>\n",
 314 |               "      <td>10889</td>\n",
 315 |               "      <td>12</td>\n",
 316 |               "    </tr>\n",
 317 |               "  </tbody>\n",
 318 |               "</table>\n",
 319 |               "</div>"
 320 |             ],
 321 |             "text/plain": [
 322 |               "                        Player  Mat   Runs  100\n",
 323 |               "0         SR Tendulkar (INDIA)  463  18426   49\n",
 324 |               "1  KC Sangakkara (Asia/ICC/SL)  404  14234   25\n",
 325 |               "2         RT Ponting (AUS/ICC)  375  13704   30\n",
 326 |               "3      ST Jayasuriya (Asia/SL)  445  13430   28\n",
 327 |               "4   DPMD Jayawardene (Asia/SL)  448  12650   19\n",
 328 |               "5              V Kohli (INDIA)  260  12311   43\n",
 329 |               "6    Inzamam-ul-Haq (Asia/PAK)  378  11739   10\n",
 330 |               "7       JH Kallis (Afr/ICC/SA)  328  11579   17\n",
 331 |               "8      SC Ganguly (Asia/INDIA)  311  11363   22\n",
 332 |               "9    R Dravid (Asia/ICC/INDIA)  344  10889   12"
 333 |             ]
 334 |           },
 335 |           "metadata": {},
 336 |           "output_type": "display_data"
 337 |         },
 338 |         {
 339 |           "name": "stdout",
 340 |           "output_type": "stream",
 341 |           "text": [
 342 |             "(92, 4)\n"
 343 |           ]
 344 |         }
 345 |       ],
 346 |       "source": [
 347 |         "col_nums = [0, 2, 5, 10]\n",
 348 |         "df_usecols_index = pd.read_csv(\"batsman_most_runs_ODI.csv\", usecols=col_nums)\n",
 349 |         "\n",
 350 |         "display(df_usecols_index.head(10))\n",
 351 |         "print(df_usecols_index.shape)"
 352 |       ]
 353 |     },
 354 |     {
 355 |       "cell_type": "markdown",
 356 |       "metadata": {
 357 |         "id": "fHdNp7hbdn_2"
 358 |       },
 359 |       "source": [
 360 |         "#### Reading only the first n number of rows"
 361 |       ]
 362 |     },
 363 |     {
 364 |       "cell_type": "code",
 365 |       "execution_count": null,
 366 |       "metadata": {
 367 |         "id": "OLgOGcCEdn_2",
 368 |         "outputId": "f05abd0e-dc1c-4c5a-da97-7e9c8e522cac"
 369 |       },
 370 |       "outputs": [
 371 |         {
 372 |           "data": {
 373 |             "text/html": [
 374 |               "<div>\n",
 375 |               "<style scoped>\n",
 376 |               "    .dataframe tbody tr th:only-of-type {\n",
 377 |               "        vertical-align: middle;\n",
 378 |               "    }\n",
 379 |               "\n",
 380 |               "    .dataframe tbody tr th {\n",
 381 |               "        vertical-align: top;\n",
 382 |               "    }\n",
 383 |               "\n",
 384 |               "    .dataframe thead th {\n",
 385 |               "        text-align: right;\n",
 386 |               "    }\n",
 387 |               "</style>\n",
 388 |               "<table border=\"1\" class=\"dataframe\">\n",
 389 |               "  <thead>\n",
 390 |               "    <tr style=\"text-align: right;\">\n",
 391 |               "      <th></th>\n",
 392 |               "      <th>Player</th>\n",
 393 |               "      <th>Span</th>\n",
 394 |               "      <th>Mat</th>\n",
 395 |               "      <th>Inns</th>\n",
 396 |               "      <th>NO</th>\n",
 397 |               "      <th>Runs</th>\n",
 398 |               "      <th>HS</th>\n",
 399 |               "      <th>Ave</th>\n",
 400 |               "      <th>BF</th>\n",
 401 |               "      <th>SR</th>\n",
 402 |               "      <th>100</th>\n",
 403 |               "      <th>50</th>\n",
 404 |               "      <th>0</th>\n",
 405 |               "      <th>4s</th>\n",
 406 |               "      <th>6s</th>\n",
 407 |               "    </tr>\n",
 408 |               "  </thead>\n",
 409 |               "  <tbody>\n",
 410 |               "    <tr>\n",
 411 |               "      <th>0</th>\n",
 412 |               "      <td>SR Tendulkar (INDIA)</td>\n",
 413 |               "      <td>1989-2012</td>\n",
 414 |               "      <td>463</td>\n",
 415 |               "      <td>452</td>\n",
 416 |               "      <td>41</td>\n",
 417 |               "      <td>18426</td>\n",
 418 |               "      <td>200*</td>\n",
 419 |               "      <td>44.83</td>\n",
 420 |               "      <td>21368</td>\n",
 421 |               "      <td>86.23</td>\n",
 422 |               "      <td>49</td>\n",
 423 |               "      <td>96</td>\n",
 424 |               "      <td>20</td>\n",
 425 |               "      <td>2016</td>\n",
 426 |               "      <td>195</td>\n",
 427 |               "    </tr>\n",
 428 |               "    <tr>\n",
 429 |               "      <th>1</th>\n",
 430 |               "      <td>KC Sangakkara (Asia/ICC/SL)</td>\n",
 431 |               "      <td>2000-2015</td>\n",
 432 |               "      <td>404</td>\n",
 433 |               "      <td>380</td>\n",
 434 |               "      <td>41</td>\n",
 435 |               "      <td>14234</td>\n",
 436 |               "      <td>169</td>\n",
 437 |               "      <td>41.98</td>\n",
 438 |               "      <td>18048</td>\n",
 439 |               "      <td>78.86</td>\n",
 440 |               "      <td>25</td>\n",
 441 |               "      <td>93</td>\n",
 442 |               "      <td>15</td>\n",
 443 |               "      <td>1385</td>\n",
 444 |               "      <td>88</td>\n",
 445 |               "    </tr>\n",
 446 |               "    <tr>\n",
 447 |               "      <th>2</th>\n",
 448 |               "      <td>RT Ponting (AUS/ICC)</td>\n",
 449 |               "      <td>1995-2012</td>\n",
 450 |               "      <td>375</td>\n",
 451 |               "      <td>365</td>\n",
 452 |               "      <td>39</td>\n",
 453 |               "      <td>13704</td>\n",
 454 |               "      <td>164</td>\n",
 455 |               "      <td>42.03</td>\n",
 456 |               "      <td>17046</td>\n",
 457 |               "      <td>80.39</td>\n",
 458 |               "      <td>30</td>\n",
 459 |               "      <td>82</td>\n",
 460 |               "      <td>20</td>\n",
 461 |               "      <td>1231</td>\n",
 462 |               "      <td>162</td>\n",
 463 |               "    </tr>\n",
 464 |               "    <tr>\n",
 465 |               "      <th>3</th>\n",
 466 |               "      <td>ST Jayasuriya (Asia/SL)</td>\n",
 467 |               "      <td>1989-2011</td>\n",
 468 |               "      <td>445</td>\n",
 469 |               "      <td>433</td>\n",
 470 |               "      <td>18</td>\n",
 471 |               "      <td>13430</td>\n",
 472 |               "      <td>189</td>\n",
 473 |               "      <td>32.36</td>\n",
 474 |               "      <td>14725</td>\n",
 475 |               "      <td>91.20</td>\n",
 476 |               "      <td>28</td>\n",
 477 |               "      <td>68</td>\n",
 478 |               "      <td>34</td>\n",
 479 |               "      <td>1500</td>\n",
 480 |               "      <td>270</td>\n",
 481 |               "    </tr>\n",
 482 |               "    <tr>\n",
 483 |               "      <th>4</th>\n",
 484 |               "      <td>DPMD Jayawardene (Asia/SL)</td>\n",
 485 |               "      <td>1998-2015</td>\n",
 486 |               "      <td>448</td>\n",
 487 |               "      <td>418</td>\n",
 488 |               "      <td>39</td>\n",
 489 |               "      <td>12650</td>\n",
 490 |               "      <td>144</td>\n",
 491 |               "      <td>33.37</td>\n",
 492 |               "      <td>16020</td>\n",
 493 |               "      <td>78.96</td>\n",
 494 |               "      <td>19</td>\n",
 495 |               "      <td>77</td>\n",
 496 |               "      <td>28</td>\n",
 497 |               "      <td>1119</td>\n",
 498 |               "      <td>76</td>\n",
 499 |               "    </tr>\n",
 500 |               "  </tbody>\n",
 501 |               "</table>\n",
 502 |               "</div>"
 503 |             ],
 504 |             "text/plain": [
 505 |               "                        Player       Span  Mat  Inns  NO   Runs    HS    Ave  \\\n",
 506 |               "0         SR Tendulkar (INDIA)  1989-2012  463   452  41  18426  200*  44.83   \n",
 507 |               "1  KC Sangakkara (Asia/ICC/SL)  2000-2015  404   380  41  14234   169  41.98   \n",
 508 |               "2         RT Ponting (AUS/ICC)  1995-2012  375   365  39  13704   164  42.03   \n",
 509 |               "3      ST Jayasuriya (Asia/SL)  1989-2011  445   433  18  13430   189  32.36   \n",
 510 |               "4   DPMD Jayawardene (Asia/SL)  1998-2015  448   418  39  12650   144  33.37   \n",
 511 |               "\n",
 512 |               "      BF     SR  100  50   0    4s   6s  \n",
 513 |               "0  21368  86.23   49  96  20  2016  195  \n",
 514 |               "1  18048  78.86   25  93  15  1385   88  \n",
 515 |               "2  17046  80.39   30  82  20  1231  162  \n",
 516 |               "3  14725  91.20   28  68  34  1500  270  \n",
 517 |               "4  16020  78.96   19  77  28  1119   76  "
 518 |             ]
 519 |           },
 520 |           "metadata": {},
 521 |           "output_type": "display_data"
 522 |         },
 523 |         {
 524 |           "name": "stdout",
 525 |           "output_type": "stream",
 526 |           "text": [
 527 |             "(50, 15)\n"
 528 |           ]
 529 |         }
 530 |       ],
 531 |       "source": [
 532 |         "df = pd.read_csv(\"batsman_most_runs_ODI.csv\", nrows=50)\n",
 533 |         "\n",
 534 |         "display(df.head())\n",
 535 |         "print(df.shape)"
 536 |       ]
 537 |     },
 538 |     {
 539 |       "cell_type": "code",
 540 |       "execution_count": null,
 541 |       "metadata": {
 542 |         "id": "SnWhEDbndn_2",
 543 |         "outputId": "70332403-4719-494c-bf84-72180291fe1f"
 544 |       },
 545 |       "outputs": [
 546 |         {
 547 |           "data": {
 548 |             "text/html": [
 549 |               "<div>\n",
 550 |               "<style scoped>\n",
 551 |               "    .dataframe tbody tr th:only-of-type {\n",
 552 |               "        vertical-align: middle;\n",
 553 |               "    }\n",
 554 |               "\n",
 555 |               "    .dataframe tbody tr th {\n",
 556 |               "        vertical-align: top;\n",
 557 |               "    }\n",
 558 |               "\n",
 559 |               "    .dataframe thead th {\n",
 560 |               "        text-align: right;\n",
 561 |               "    }\n",
 562 |               "</style>\n",
 563 |               "<table border=\"1\" class=\"dataframe\">\n",
 564 |               "  <thead>\n",
 565 |               "    <tr style=\"text-align: right;\">\n",
 566 |               "      <th></th>\n",
 567 |               "      <th>Player</th>\n",
 568 |               "      <th>Span</th>\n",
 569 |               "      <th>Mat</th>\n",
 570 |               "      <th>Inns</th>\n",
 571 |               "      <th>NO</th>\n",
 572 |               "      <th>Runs</th>\n",
 573 |               "      <th>HS</th>\n",
 574 |               "      <th>Ave</th>\n",
 575 |               "      <th>BF</th>\n",
 576 |               "      <th>SR</th>\n",
 577 |               "      <th>100</th>\n",
 578 |               "      <th>50</th>\n",
 579 |               "      <th>0</th>\n",
 580 |               "      <th>4s</th>\n",
 581 |               "      <th>6s</th>\n",
 582 |               "    </tr>\n",
 583 |               "  </thead>\n",
 584 |               "  <tbody>\n",
 585 |               "    <tr>\n",
 586 |               "      <th>44</th>\n",
 587 |               "      <td>MJ Guptill (NZ)</td>\n",
 588 |               "      <td>2009-2021</td>\n",
 589 |               "      <td>186</td>\n",
 590 |               "      <td>183</td>\n",
 591 |               "      <td>19</td>\n",
 592 |               "      <td>6927</td>\n",
 593 |               "      <td>237*</td>\n",
 594 |               "      <td>42.23</td>\n",
 595 |               "      <td>7896</td>\n",
 596 |               "      <td>87.72</td>\n",
 597 |               "      <td>16</td>\n",
 598 |               "      <td>37</td>\n",
 599 |               "      <td>15</td>\n",
 600 |               "      <td>702</td>\n",
 601 |               "      <td>181</td>\n",
 602 |               "    </tr>\n",
 603 |               "    <tr>\n",
 604 |               "      <th>2</th>\n",
 605 |               "      <td>RT Ponting (AUS/ICC)</td>\n",
 606 |               "      <td>1995-2012</td>\n",
 607 |               "      <td>375</td>\n",
 608 |               "      <td>365</td>\n",
 609 |               "      <td>39</td>\n",
 610 |               "      <td>13704</td>\n",
 611 |               "      <td>164</td>\n",
 612 |               "      <td>42.03</td>\n",
 613 |               "      <td>17046</td>\n",
 614 |               "      <td>80.39</td>\n",
 615 |               "      <td>30</td>\n",
 616 |               "      <td>82</td>\n",
 617 |               "      <td>20</td>\n",
 618 |               "      <td>1231</td>\n",
 619 |               "      <td>162</td>\n",
 620 |               "    </tr>\n",
 621 |               "  </tbody>\n",
 622 |               "</table>\n",
 623 |               "</div>"
 624 |             ],
 625 |             "text/plain": [
 626 |               "                  Player       Span  Mat  Inns  NO   Runs    HS    Ave     BF  \\\n",
 627 |               "44       MJ Guptill (NZ)  2009-2021  186   183  19   6927  237*  42.23   7896   \n",
 628 |               "2   RT Ponting (AUS/ICC)  1995-2012  375   365  39  13704   164  42.03  17046   \n",
 629 |               "\n",
 630 |               "       SR  100  50   0    4s   6s  \n",
 631 |               "44  87.72   16  37  15   702  181  \n",
 632 |               "2   80.39   30  82  20  1231  162  "
 633 |             ]
 634 |           },
 635 |           "execution_count": 220,
 636 |           "metadata": {},
 637 |           "output_type": "execute_result"
 638 |         }
 639 |       ],
 640 |       "source": [
 641 |         "# showing randomly 2 different rows\n",
 642 |         "df.sample(2)"
 643 |       ]
 644 |     },
 645 |     {
 646 |       "cell_type": "markdown",
 647 |       "metadata": {
 648 |         "id": "irul-Zt_dn_3"
 649 |       },
 650 |       "source": [
 651 |         "#### Reading Excel file"
 652 |       ]
 653 |     },
 654 |     {
 655 |       "cell_type": "code",
 656 |       "execution_count": null,
 657 |       "metadata": {
 658 |         "id": "AcNg73d-dn_3",
 659 |         "outputId": "09e29b8d-eff4-43a6-e42c-4a1397068304"
 660 |       },
 661 |       "outputs": [
 662 |         {
 663 |           "data": {
 664 |             "text/html": [
 665 |               "<div>\n",
 666 |               "<style scoped>\n",
 667 |               "    .dataframe tbody tr th:only-of-type {\n",
 668 |               "        vertical-align: middle;\n",
 669 |               "    }\n",
 670 |               "\n",
 671 |               "    .dataframe tbody tr th {\n",
 672 |               "        vertical-align: top;\n",
 673 |               "    }\n",
 674 |               "\n",
 675 |               "    .dataframe thead th {\n",
 676 |               "        text-align: right;\n",
 677 |               "    }\n",
 678 |               "</style>\n",
 679 |               "<table border=\"1\" class=\"dataframe\">\n",
 680 |               "  <thead>\n",
 681 |               "    <tr style=\"text-align: right;\">\n",
 682 |               "      <th></th>\n",
 683 |               "      <th>Player</th>\n",
 684 |               "      <th>Span</th>\n",
 685 |               "      <th>Mat</th>\n",
 686 |               "      <th>Inns</th>\n",
 687 |               "      <th>NO</th>\n",
 688 |               "      <th>Runs</th>\n",
 689 |               "      <th>HS</th>\n",
 690 |               "      <th>Ave</th>\n",
 691 |               "      <th>BF</th>\n",
 692 |               "      <th>SR</th>\n",
 693 |               "      <th>100</th>\n",
 694 |               "      <th>50</th>\n",
 695 |               "      <th>0</th>\n",
 696 |               "      <th>4s</th>\n",
 697 |               "      <th>6s</th>\n",
 698 |               "    </tr>\n",
 699 |               "  </thead>\n",
 700 |               "  <tbody>\n",
 701 |               "    <tr>\n",
 702 |               "      <th>0</th>\n",
 703 |               "      <td>SR Tendulkar (INDIA)</td>\n",
 704 |               "      <td>1989-2012</td>\n",
 705 |               "      <td>463</td>\n",
 706 |               "      <td>452</td>\n",
 707 |               "      <td>41</td>\n",
 708 |               "      <td>18426</td>\n",
 709 |               "      <td>200*</td>\n",
 710 |               "      <td>44.83</td>\n",
 711 |               "      <td>21368</td>\n",
 712 |               "      <td>86.23</td>\n",
 713 |               "      <td>49</td>\n",
 714 |               "      <td>96</td>\n",
 715 |               "      <td>20</td>\n",
 716 |               "      <td>2016</td>\n",
 717 |               "      <td>195</td>\n",
 718 |               "    </tr>\n",
 719 |               "    <tr>\n",
 720 |               "      <th>1</th>\n",
 721 |               "      <td>KC Sangakkara (Asia/ICC/SL)</td>\n",
 722 |               "      <td>2000-2015</td>\n",
 723 |               "      <td>404</td>\n",
 724 |               "      <td>380</td>\n",
 725 |               "      <td>41</td>\n",
 726 |               "      <td>14234</td>\n",
 727 |               "      <td>169</td>\n",
 728 |               "      <td>41.98</td>\n",
 729 |               "      <td>18048</td>\n",
 730 |               "      <td>78.86</td>\n",
 731 |               "      <td>25</td>\n",
 732 |               "      <td>93</td>\n",
 733 |               "      <td>15</td>\n",
 734 |               "      <td>1385</td>\n",
 735 |               "      <td>88</td>\n",
 736 |               "    </tr>\n",
 737 |               "    <tr>\n",
 738 |               "      <th>2</th>\n",
 739 |               "      <td>RT Ponting (AUS/ICC)</td>\n",
 740 |               "      <td>1995-2012</td>\n",
 741 |               "      <td>375</td>\n",
 742 |               "      <td>365</td>\n",
 743 |               "      <td>39</td>\n",
 744 |               "      <td>13704</td>\n",
 745 |               "      <td>164</td>\n",
 746 |               "      <td>42.03</td>\n",
 747 |               "      <td>17046</td>\n",
 748 |               "      <td>80.39</td>\n",
 749 |               "      <td>30</td>\n",
 750 |               "      <td>82</td>\n",
 751 |               "      <td>20</td>\n",
 752 |               "      <td>1231</td>\n",
 753 |               "      <td>162</td>\n",
 754 |               "    </tr>\n",
 755 |               "    <tr>\n",
 756 |               "      <th>3</th>\n",
 757 |               "      <td>ST Jayasuriya (Asia/SL)</td>\n",
 758 |               "      <td>1989-2011</td>\n",
 759 |               "      <td>445</td>\n",
 760 |               "      <td>433</td>\n",
 761 |               "      <td>18</td>\n",
 762 |               "      <td>13430</td>\n",
 763 |               "      <td>189</td>\n",
 764 |               "      <td>32.36</td>\n",
 765 |               "      <td>14725</td>\n",
 766 |               "      <td>91.20</td>\n",
 767 |               "      <td>28</td>\n",
 768 |               "      <td>68</td>\n",
 769 |               "      <td>34</td>\n",
 770 |               "      <td>1500</td>\n",
 771 |               "      <td>270</td>\n",
 772 |               "    </tr>\n",
 773 |               "    <tr>\n",
 774 |               "      <th>4</th>\n",
 775 |               "      <td>DPMD Jayawardene (Asia/SL)</td>\n",
 776 |               "      <td>1998-2015</td>\n",
 777 |               "      <td>448</td>\n",
 778 |               "      <td>418</td>\n",
 779 |               "      <td>39</td>\n",
 780 |               "      <td>12650</td>\n",
 781 |               "      <td>144</td>\n",
 782 |               "      <td>33.37</td>\n",
 783 |               "      <td>16020</td>\n",
 784 |               "      <td>78.96</td>\n",
 785 |               "      <td>19</td>\n",
 786 |               "      <td>77</td>\n",
 787 |               "      <td>28</td>\n",
 788 |               "      <td>1119</td>\n",
 789 |               "      <td>76</td>\n",
 790 |               "    </tr>\n",
 791 |               "  </tbody>\n",
 792 |               "</table>\n",
 793 |               "</div>"
 794 |             ],
 795 |             "text/plain": [
 796 |               "                        Player       Span  Mat  Inns  NO   Runs    HS    Ave  \\\n",
 797 |               "0         SR Tendulkar (INDIA)  1989-2012  463   452  41  18426  200*  44.83   \n",
 798 |               "1  KC Sangakkara (Asia/ICC/SL)  2000-2015  404   380  41  14234   169  41.98   \n",
 799 |               "2         RT Ponting (AUS/ICC)  1995-2012  375   365  39  13704   164  42.03   \n",
 800 |               "3      ST Jayasuriya (Asia/SL)  1989-2011  445   433  18  13430   189  32.36   \n",
 801 |               "4   DPMD Jayawardene (Asia/SL)  1998-2015  448   418  39  12650   144  33.37   \n",
 802 |               "\n",
 803 |               "      BF     SR  100  50   0    4s   6s  \n",
 804 |               "0  21368  86.23   49  96  20  2016  195  \n",
 805 |               "1  18048  78.86   25  93  15  1385   88  \n",
 806 |               "2  17046  80.39   30  82  20  1231  162  \n",
 807 |               "3  14725  91.20   28  68  34  1500  270  \n",
 808 |               "4  16020  78.96   19  77  28  1119   76  "
 809 |             ]
 810 |           },
 811 |           "metadata": {},
 812 |           "output_type": "display_data"
 813 |         }
 814 |       ],
 815 |       "source": [
 816 |         "df = pd.read_excel(\"ODI_cricket.xlsx\", sheet_name=\"batsman\", engine=\"openpyxl\")\n",
 817 |         "\n",
 818 |         "display(df.head())"
 819 |       ]
 820 |     },
 821 |     {
 822 |       "cell_type": "code",
 823 |       "execution_count": null,
 824 |       "metadata": {
 825 |         "id": "2XfuJLPudn_1"
 826 |       },
 827 |       "outputs": [],
 828 |       "source": [
 829 |         "# print(df.shape)\n",
 830 |         "# col_names = ['Player', 'Mat', 'Runs', 'SR']\n",
 831 |         "# # selecting columns after data importing\n",
 832 |         "# df = df[col_names]\n",
 833 |         "\n",
 834 |         "# print(df.shape)"
 835 |       ]
 836 |     },
 837 |     {
 838 |       "cell_type": "markdown",
 839 |       "metadata": {
 840 |         "id": "Y8k5_uuIdn_3"
 841 |       },
 842 |       "source": [
 843 |         "#### How to rename the column names?"
 844 |       ]
 845 |     },
 846 |     {
 847 |       "cell_type": "code",
 848 |       "execution_count": null,
 849 |       "metadata": {
 850 |         "id": "ZxXbqzobv8hq",
 851 |         "outputId": "ff6df881-baf0-4e60-a7dc-56c889fc1f1c"
 852 |       },
 853 |       "outputs": [
 854 |         {
 855 |           "name": "stdout",
 856 |           "output_type": "stream",
 857 |           "text": [
 858 |             "Index(['Player',   'Span',    'Mat',   'Inns',     'NO',   'Runs',     'HS',\n",
 859 |             "          'Ave',     'BF',     'SR',      100,       50,        0,     '4s',\n",
 860 |             "           '6s'],\n",
 861 |             "      dtype='object')\n"
 862 |           ]
 863 |         }
 864 |       ],
 865 |       "source": [
 866 |         "print(df.columns)"
 867 |       ]
 868 |     },
 869 |     {
 870 |       "cell_type": "code",
 871 |       "execution_count": null,
 872 |       "metadata": {
 873 |         "id": "pY76ppQbv8hq",
 874 |         "outputId": "e96f5f7a-95d7-4e55-9370-a793bd0fb99e"
 875 |       },
 876 |       "outputs": [
 877 |         {
 878 |           "data": {
 879 |             "text/html": [
 880 |               "<div>\n",
 881 |               "<style scoped>\n",
 882 |               "    .dataframe tbody tr th:only-of-type {\n",
 883 |               "        vertical-align: middle;\n",
 884 |               "    }\n",
 885 |               "\n",
 886 |               "    .dataframe tbody tr th {\n",
 887 |               "        vertical-align: top;\n",
 888 |               "    }\n",
 889 |               "\n",
 890 |               "    .dataframe thead th {\n",
 891 |               "        text-align: right;\n",
 892 |               "    }\n",
 893 |               "</style>\n",
 894 |               "<table border=\"1\" class=\"dataframe\">\n",
 895 |               "  <thead>\n",
 896 |               "    <tr style=\"text-align: right;\">\n",
 897 |               "      <th></th>\n",
 898 |               "      <th>Player</th>\n",
 899 |               "      <th>Span</th>\n",
 900 |               "      <th>Match</th>\n",
 901 |               "      <th>Innings</th>\n",
 902 |               "      <th>NotOut</th>\n",
 903 |               "      <th>Runs</th>\n",
 904 |               "      <th>Highest_score</th>\n",
 905 |               "      <th>Average</th>\n",
 906 |               "      <th>Balls_Faced</th>\n",
 907 |               "      <th>Strike_Rate</th>\n",
 908 |               "      <th>Centuries</th>\n",
 909 |               "      <th>Half_centuries</th>\n",
 910 |               "      <th>Ducks</th>\n",
 911 |               "      <th>Fours</th>\n",
 912 |               "      <th>Sixes</th>\n",
 913 |               "    </tr>\n",
 914 |               "  </thead>\n",
 915 |               "  <tbody>\n",
 916 |               "    <tr>\n",
 917 |               "      <th>0</th>\n",
 918 |               "      <td>SR Tendulkar (INDIA)</td>\n",
 919 |               "      <td>1989-2012</td>\n",
 920 |               "      <td>463</td>\n",
 921 |               "      <td>452</td>\n",
 922 |               "      <td>41</td>\n",
 923 |               "      <td>18426</td>\n",
 924 |               "      <td>200*</td>\n",
 925 |               "      <td>44.83</td>\n",
 926 |               "      <td>21368</td>\n",
 927 |               "      <td>86.23</td>\n",
 928 |               "      <td>49</td>\n",
 929 |               "      <td>96</td>\n",
 930 |               "      <td>20</td>\n",
 931 |               "      <td>2016</td>\n",
 932 |               "      <td>195</td>\n",
 933 |               "    </tr>\n",
 934 |               "    <tr>\n",
 935 |               "      <th>1</th>\n",
 936 |               "      <td>KC Sangakkara (Asia/ICC/SL)</td>\n",
 937 |               "      <td>2000-2015</td>\n",
 938 |               "      <td>404</td>\n",
 939 |               "      <td>380</td>\n",
 940 |               "      <td>41</td>\n",
 941 |               "      <td>14234</td>\n",
 942 |               "      <td>169</td>\n",
 943 |               "      <td>41.98</td>\n",
 944 |               "      <td>18048</td>\n",
 945 |               "      <td>78.86</td>\n",
 946 |               "      <td>25</td>\n",
 947 |               "      <td>93</td>\n",
 948 |               "      <td>15</td>\n",
 949 |               "      <td>1385</td>\n",
 950 |               "      <td>88</td>\n",
 951 |               "    </tr>\n",
 952 |               "    <tr>\n",
 953 |               "      <th>2</th>\n",
 954 |               "      <td>RT Ponting (AUS/ICC)</td>\n",
 955 |               "      <td>1995-2012</td>\n",
 956 |               "      <td>375</td>\n",
 957 |               "      <td>365</td>\n",
 958 |               "      <td>39</td>\n",
 959 |               "      <td>13704</td>\n",
 960 |               "      <td>164</td>\n",
 961 |               "      <td>42.03</td>\n",
 962 |               "      <td>17046</td>\n",
 963 |               "      <td>80.39</td>\n",
 964 |               "      <td>30</td>\n",
 965 |               "      <td>82</td>\n",
 966 |               "      <td>20</td>\n",
 967 |               "      <td>1231</td>\n",
 968 |               "      <td>162</td>\n",
 969 |               "    </tr>\n",
 970 |               "    <tr>\n",
 971 |               "      <th>3</th>\n",
 972 |               "      <td>ST Jayasuriya (Asia/SL)</td>\n",
 973 |               "      <td>1989-2011</td>\n",
 974 |               "      <td>445</td>\n",
 975 |               "      <td>433</td>\n",
 976 |               "      <td>18</td>\n",
 977 |               "      <td>13430</td>\n",
 978 |               "      <td>189</td>\n",
 979 |               "      <td>32.36</td>\n",
 980 |               "      <td>14725</td>\n",
 981 |               "      <td>91.20</td>\n",
 982 |               "      <td>28</td>\n",
 983 |               "      <td>68</td>\n",
 984 |               "      <td>34</td>\n",
 985 |               "      <td>1500</td>\n",
 986 |               "      <td>270</td>\n",
 987 |               "    </tr>\n",
 988 |               "    <tr>\n",
 989 |               "      <th>4</th>\n",
 990 |               "      <td>DPMD Jayawardene (Asia/SL)</td>\n",
 991 |               "      <td>1998-2015</td>\n",
 992 |               "      <td>448</td>\n",
 993 |               "      <td>418</td>\n",
 994 |               "      <td>39</td>\n",
 995 |               "      <td>12650</td>\n",
 996 |               "      <td>144</td>\n",
 997 |               "      <td>33.37</td>\n",
 998 |               "      <td>16020</td>\n",
 999 |               "      <td>78.96</td>\n",
1000 |               "      <td>19</td>\n",
1001 |               "      <td>77</td>\n",
1002 |               "      <td>28</td>\n",
1003 |               "      <td>1119</td>\n",
1004 |               "      <td>76</td>\n",
1005 |               "    </tr>\n",
1006 |               "  </tbody>\n",
1007 |               "</table>\n",
1008 |               "</div>"
1009 |             ],
1010 |             "text/plain": [
1011 |               "                        Player       Span  Match  Innings  NotOut   Runs  \\\n",
1012 |               "0         SR Tendulkar (INDIA)  1989-2012    463      452      41  18426   \n",
1013 |               "1  KC Sangakkara (Asia/ICC/SL)  2000-2015    404      380      41  14234   \n",
1014 |               "2         RT Ponting (AUS/ICC)  1995-2012    375      365      39  13704   \n",
1015 |               "3      ST Jayasuriya (Asia/SL)  1989-2011    445      433      18  13430   \n",
1016 |               "4   DPMD Jayawardene (Asia/SL)  1998-2015    448      418      39  12650   \n",
1017 |               "\n",
1018 |               "  Highest_score  Average  Balls_Faced  Strike_Rate  Centuries  Half_centuries  \\\n",
1019 |               "0          200*    44.83        21368        86.23         49              96   \n",
1020 |               "1           169    41.98        18048        78.86         25              93   \n",
1021 |               "2           164    42.03        17046        80.39         30              82   \n",
1022 |               "3           189    32.36        14725        91.20         28              68   \n",
1023 |               "4           144    33.37        16020        78.96         19              77   \n",
1024 |               "\n",
1025 |               "   Ducks Fours Sixes  \n",
1026 |               "0     20  2016   195  \n",
1027 |               "1     15  1385    88  \n",
1028 |               "2     20  1231   162  \n",
1029 |               "3     34  1500   270  \n",
1030 |               "4     28  1119    76  "
1031 |             ]
1032 |           },
1033 |           "metadata": {},
1034 |           "output_type": "display_data"
1035 |         }
1036 |       ],
1037 |       "source": [
1038 |         "df = df.rename(columns={'Mat':'Match', \n",
1039 |         "                        'Inns':'Innings',\n",
1040 |         "                        'NO': 'NotOut',\n",
1041 |         "                        'HS': 'Highest_score',\n",
1042 |         "                        'Ave': 'Average',\n",
1043 |         "                        'BF': 'Balls_Faced',\n",
1044 |         "                        'SR': 'Strike_Rate',\n",
1045 |         "                        100: 'Centuries',\n",
1046 |         "                        50: 'Half_centuries',\n",
1047 |         "                        0: 'Ducks',\n",
1048 |         "                        \"4s\": \"Fours\",\n",
1049 |         "                        \"6s\": \"Sixes\"})\n",
1050 |         "\n",
1051 |         "display(df.head())"
1052 |       ]
1053 |     },
1054 |     {
1055 |       "cell_type": "markdown",
1056 |       "metadata": {
1057 |         "id": "V8AUroJcv8hr"
1058 |       },
1059 |       "source": [
1060 |         "#### How to split a column and create two new columns?"
1061 |       ]
1062 |     },
1063 |     {
1064 |       "cell_type": "code",
1065 |       "execution_count": null,
1066 |       "metadata": {
1067 |         "id": "GllvKT30v8hr",
1068 |         "outputId": "9e5444b7-23fe-4462-86b9-6dcf1fd84f95"
1069 |       },
1070 |       "outputs": [
1071 |         {
1072 |           "data": {
1073 |             "text/html": [
1074 |               "<div>\n",
1075 |               "<style scoped>\n",
1076 |               "    .dataframe tbody tr th:only-of-type {\n",
1077 |               "        vertical-align: middle;\n",
1078 |               "    }\n",
1079 |               "\n",
1080 |               "    .dataframe tbody tr th {\n",
1081 |               "        vertical-align: top;\n",
1082 |               "    }\n",
1083 |               "\n",
1084 |               "    .dataframe thead th {\n",
1085 |               "        text-align: right;\n",
1086 |               "    }\n",
1087 |               "</style>\n",
1088 |               "<table border=\"1\" class=\"dataframe\">\n",
1089 |               "  <thead>\n",
1090 |               "    <tr style=\"text-align: right;\">\n",
1091 |               "      <th></th>\n",
1092 |               "      <th>0</th>\n",
1093 |               "      <th>1</th>\n",
1094 |               "    </tr>\n",
1095 |               "  </thead>\n",
1096 |               "  <tbody>\n",
1097 |               "    <tr>\n",
1098 |               "      <th>0</th>\n",
1099 |               "      <td>SR Tendulkar</td>\n",
1100 |               "      <td>INDIA)</td>\n",
1101 |               "    </tr>\n",
1102 |               "    <tr>\n",
1103 |               "      <th>1</th>\n",
1104 |               "      <td>KC Sangakkara</td>\n",
1105 |               "      <td>Asia/ICC/SL)</td>\n",
1106 |               "    </tr>\n",
1107 |               "    <tr>\n",
1108 |               "      <th>2</th>\n",
1109 |               "      <td>RT Ponting</td>\n",
1110 |               "      <td>AUS/ICC)</td>\n",
1111 |               "    </tr>\n",
1112 |               "    <tr>\n",
1113 |               "      <th>3</th>\n",
1114 |               "      <td>ST Jayasuriya</td>\n",
1115 |               "      <td>Asia/SL)</td>\n",
1116 |               "    </tr>\n",
1117 |               "    <tr>\n",
1118 |               "      <th>4</th>\n",
1119 |               "      <td>DPMD Jayawardene</td>\n",
1120 |               "      <td>Asia/SL)</td>\n",
1121 |               "    </tr>\n",
1122 |               "    <tr>\n",
1123 |               "      <th>5</th>\n",
1124 |               "      <td>V Kohli</td>\n",
1125 |               "      <td>INDIA)</td>\n",
1126 |               "    </tr>\n",
1127 |               "    <tr>\n",
1128 |               "      <th>6</th>\n",
1129 |               "      <td>Inzamam-ul-Haq</td>\n",
1130 |               "      <td>Asia/PAK)</td>\n",
1131 |               "    </tr>\n",
1132 |               "    <tr>\n",
1133 |               "      <th>7</th>\n",
1134 |               "      <td>JH Kallis</td>\n",
1135 |               "      <td>Afr/ICC/SA)</td>\n",
1136 |               "    </tr>\n",
1137 |               "    <tr>\n",
1138 |               "      <th>8</th>\n",
1139 |               "      <td>SC Ganguly</td>\n",
1140 |               "      <td>Asia/INDIA)</td>\n",
1141 |               "    </tr>\n",
1142 |               "    <tr>\n",
1143 |               "      <th>9</th>\n",
1144 |               "      <td>R Dravid</td>\n",
1145 |               "      <td>Asia/ICC/INDIA)</td>\n",
1146 |               "    </tr>\n",
1147 |               "  </tbody>\n",
1148 |               "</table>\n",
1149 |               "</div>"
1150 |             ],
1151 |             "text/plain": [
1152 |               "                   0                1\n",
1153 |               "0      SR Tendulkar            INDIA)\n",
1154 |               "1     KC Sangakkara      Asia/ICC/SL)\n",
1155 |               "2        RT Ponting          AUS/ICC)\n",
1156 |               "3     ST Jayasuriya          Asia/SL)\n",
1157 |               "4  DPMD Jayawardene          Asia/SL)\n",
1158 |               "5           V Kohli            INDIA)\n",
1159 |               "6    Inzamam-ul-Haq         Asia/PAK)\n",
1160 |               "7         JH Kallis       Afr/ICC/SA)\n",
1161 |               "8        SC Ganguly       Asia/INDIA)\n",
1162 |               "9          R Dravid   Asia/ICC/INDIA)"
1163 |             ]
1164 |           },
1165 |           "metadata": {},
1166 |           "output_type": "display_data"
1167 |         }
1168 |       ],
1169 |       "source": [
1170 |         "df_player = df['Player'].str.split(\"(\", expand=True)\n",
1171 |         "\n",
1172 |         "display(df_player.head(10))"
1173 |       ]
1174 |     },
1175 |     {
1176 |       "cell_type": "code",
1177 |       "execution_count": null,
1178 |       "metadata": {
1179 |         "id": "pIFx34ijv8hr",
1180 |         "outputId": "fda321ee-a63a-42a6-cdb3-645920849483"
1181 |       },
1182 |       "outputs": [
1183 |         {
1184 |           "data": {
1185 |             "text/html": [
1186 |               "<div>\n",
1187 |               "<style scoped>\n",
1188 |               "    .dataframe tbody tr th:only-of-type {\n",
1189 |               "        vertical-align: middle;\n",
1190 |               "    }\n",
1191 |               "\n",
1192 |               "    .dataframe tbody tr th {\n",
1193 |               "        vertical-align: top;\n",
1194 |               "    }\n",
1195 |               "\n",
1196 |               "    .dataframe thead th {\n",
1197 |               "        text-align: right;\n",
1198 |               "    }\n",
1199 |               "</style>\n",
1200 |               "<table border=\"1\" class=\"dataframe\">\n",
1201 |               "  <thead>\n",
1202 |               "    <tr style=\"text-align: right;\">\n",
1203 |               "      <th></th>\n",
1204 |               "      <th>Player</th>\n",
1205 |               "      <th>Span</th>\n",
1206 |               "      <th>Match</th>\n",
1207 |               "      <th>Innings</th>\n",
1208 |               "      <th>NotOut</th>\n",
1209 |               "      <th>Runs</th>\n",
1210 |               "      <th>Highest_score</th>\n",
1211 |               "      <th>Average</th>\n",
1212 |               "      <th>Balls_Faced</th>\n",
1213 |               "      <th>Strike_Rate</th>\n",
1214 |               "      <th>Centuries</th>\n",
1215 |               "      <th>Half_centuries</th>\n",
1216 |               "      <th>Ducks</th>\n",
1217 |               "      <th>Fours</th>\n",
1218 |               "      <th>Sixes</th>\n",
1219 |               "      <th>Player_Name</th>\n",
1220 |               "      <th>Country</th>\n",
1221 |               "    </tr>\n",
1222 |               "  </thead>\n",
1223 |               "  <tbody>\n",
1224 |               "    <tr>\n",
1225 |               "      <th>0</th>\n",
1226 |               "      <td>SR Tendulkar (INDIA)</td>\n",
1227 |               "      <td>1989-2012</td>\n",
1228 |               "      <td>463</td>\n",
1229 |               "      <td>452</td>\n",
1230 |               "      <td>41</td>\n",
1231 |               "      <td>18426</td>\n",
1232 |               "      <td>200*</td>\n",
1233 |               "      <td>44.83</td>\n",
1234 |               "      <td>21368</td>\n",
1235 |               "      <td>86.23</td>\n",
1236 |               "      <td>49</td>\n",
1237 |               "      <td>96</td>\n",
1238 |               "      <td>20</td>\n",
1239 |               "      <td>2016</td>\n",
1240 |               "      <td>195</td>\n",
1241 |               "      <td>SR Tendulkar</td>\n",
1242 |               "      <td>INDIA)</td>\n",
1243 |               "    </tr>\n",
1244 |               "    <tr>\n",
1245 |               "      <th>1</th>\n",
1246 |               "      <td>KC Sangakkara (Asia/ICC/SL)</td>\n",
1247 |               "      <td>2000-2015</td>\n",
1248 |               "      <td>404</td>\n",
1249 |               "      <td>380</td>\n",
1250 |               "      <td>41</td>\n",
1251 |               "      <td>14234</td>\n",
1252 |               "      <td>169</td>\n",
1253 |               "      <td>41.98</td>\n",
1254 |               "      <td>18048</td>\n",
1255 |               "      <td>78.86</td>\n",
1256 |               "      <td>25</td>\n",
1257 |               "      <td>93</td>\n",
1258 |               "      <td>15</td>\n",
1259 |               "      <td>1385</td>\n",
1260 |               "      <td>88</td>\n",
1261 |               "      <td>KC Sangakkara</td>\n",
1262 |               "      <td>Asia/ICC/SL)</td>\n",
1263 |               "    </tr>\n",
1264 |               "    <tr>\n",
1265 |               "      <th>2</th>\n",
1266 |               "      <td>RT Ponting (AUS/ICC)</td>\n",
1267 |               "      <td>1995-2012</td>\n",
1268 |               "      <td>375</td>\n",
1269 |               "      <td>365</td>\n",
1270 |               "      <td>39</td>\n",
1271 |               "      <td>13704</td>\n",
1272 |               "      <td>164</td>\n",
1273 |               "      <td>42.03</td>\n",
1274 |               "      <td>17046</td>\n",
1275 |               "      <td>80.39</td>\n",
1276 |               "      <td>30</td>\n",
1277 |               "      <td>82</td>\n",
1278 |               "      <td>20</td>\n",
1279 |               "      <td>1231</td>\n",
1280 |               "      <td>162</td>\n",
1281 |               "      <td>RT Ponting</td>\n",
1282 |               "      <td>AUS/ICC)</td>\n",
1283 |               "    </tr>\n",
1284 |               "    <tr>\n",
1285 |               "      <th>3</th>\n",
1286 |               "      <td>ST Jayasuriya (Asia/SL)</td>\n",
1287 |               "      <td>1989-2011</td>\n",
1288 |               "      <td>445</td>\n",
1289 |               "      <td>433</td>\n",
1290 |               "      <td>18</td>\n",
1291 |               "      <td>13430</td>\n",
1292 |               "      <td>189</td>\n",
1293 |               "      <td>32.36</td>\n",
1294 |               "      <td>14725</td>\n",
1295 |               "      <td>91.20</td>\n",
1296 |               "      <td>28</td>\n",
1297 |               "      <td>68</td>\n",
1298 |               "      <td>34</td>\n",
1299 |               "      <td>1500</td>\n",
1300 |               "      <td>270</td>\n",
1301 |               "      <td>ST Jayasuriya</td>\n",
1302 |               "      <td>Asia/SL)</td>\n",
1303 |               "    </tr>\n",
1304 |               "    <tr>\n",
1305 |               "      <th>4</th>\n",
1306 |               "      <td>DPMD Jayawardene (Asia/SL)</td>\n",
1307 |               "      <td>1998-2015</td>\n",
1308 |               "      <td>448</td>\n",
1309 |               "      <td>418</td>\n",
1310 |               "      <td>39</td>\n",
1311 |               "      <td>12650</td>\n",
1312 |               "      <td>144</td>\n",
1313 |               "      <td>33.37</td>\n",
1314 |               "      <td>16020</td>\n",
1315 |               "      <td>78.96</td>\n",
1316 |               "      <td>19</td>\n",
1317 |               "      <td>77</td>\n",
1318 |               "      <td>28</td>\n",
1319 |               "      <td>1119</td>\n",
1320 |               "      <td>76</td>\n",
1321 |               "      <td>DPMD Jayawardene</td>\n",
1322 |               "      <td>Asia/SL)</td>\n",
1323 |               "    </tr>\n",
1324 |               "  </tbody>\n",
1325 |               "</table>\n",
1326 |               "</div>"
1327 |             ],
1328 |             "text/plain": [
1329 |               "                        Player       Span  Match  Innings  NotOut   Runs  \\\n",
1330 |               "0         SR Tendulkar (INDIA)  1989-2012    463      452      41  18426   \n",
1331 |               "1  KC Sangakkara (Asia/ICC/SL)  2000-2015    404      380      41  14234   \n",
1332 |               "2         RT Ponting (AUS/ICC)  1995-2012    375      365      39  13704   \n",
1333 |               "3      ST Jayasuriya (Asia/SL)  1989-2011    445      433      18  13430   \n",
1334 |               "4   DPMD Jayawardene (Asia/SL)  1998-2015    448      418      39  12650   \n",
1335 |               "\n",
1336 |               "  Highest_score  Average  Balls_Faced  Strike_Rate  Centuries  Half_centuries  \\\n",
1337 |               "0          200*    44.83        21368        86.23         49              96   \n",
1338 |               "1           169    41.98        18048        78.86         25              93   \n",
1339 |               "2           164    42.03        17046        80.39         30              82   \n",
1340 |               "3           189    32.36        14725        91.20         28              68   \n",
1341 |               "4           144    33.37        16020        78.96         19              77   \n",
1342 |               "\n",
1343 |               "   Ducks Fours Sixes        Player_Name       Country  \n",
1344 |               "0     20  2016   195      SR Tendulkar         INDIA)  \n",
1345 |               "1     15  1385    88     KC Sangakkara   Asia/ICC/SL)  \n",
1346 |               "2     20  1231   162        RT Ponting       AUS/ICC)  \n",
1347 |               "3     34  1500   270     ST Jayasuriya       Asia/SL)  \n",
1348 |               "4     28  1119    76  DPMD Jayawardene       Asia/SL)  "
1349 |             ]
1350 |           },
1351 |           "metadata": {},
1352 |           "output_type": "display_data"
1353 |         }
1354 |       ],
1355 |       "source": [
1356 |         "df[[\"Player_Name\", \"Country\"]] = df['Player'].str.split(\"(\", expand=True)\n",
1357 |         "\n",
1358 |         "display(df.head())"
1359 |       ]
1360 |     },
1361 |     {
1362 |       "cell_type": "markdown",
1363 |       "metadata": {
1364 |         "id": "wa8kloXJv8hs"
1365 |       },
1366 |       "source": [
1367 |         "#### How to remove a column?"
1368 |       ]
1369 |     },
1370 |     {
1371 |       "cell_type": "code",
1372 |       "execution_count": null,
1373 |       "metadata": {
1374 |         "id": "z9VLMvGOv8hs",
1375 |         "outputId": "a6f57a70-56b4-4380-93c4-a85001925700"
1376 |       },
1377 |       "outputs": [
1378 |         {
1379 |           "data": {
1380 |             "text/html": [
1381 |               "<div>\n",
1382 |               "<style scoped>\n",
1383 |               "    .dataframe tbody tr th:only-of-type {\n",
1384 |               "        vertical-align: middle;\n",
1385 |               "    }\n",
1386 |               "\n",
1387 |               "    .dataframe tbody tr th {\n",
1388 |               "        vertical-align: top;\n",
1389 |               "    }\n",
1390 |               "\n",
1391 |               "    .dataframe thead th {\n",
1392 |               "        text-align: right;\n",
1393 |               "    }\n",
1394 |               "</style>\n",
1395 |               "<table border=\"1\" class=\"dataframe\">\n",
1396 |               "  <thead>\n",
1397 |               "    <tr style=\"text-align: right;\">\n",
1398 |               "      <th></th>\n",
1399 |               "      <th>Span</th>\n",
1400 |               "      <th>Match</th>\n",
1401 |               "      <th>Innings</th>\n",
1402 |               "      <th>NotOut</th>\n",
1403 |               "      <th>Runs</th>\n",
1404 |               "      <th>Highest_score</th>\n",
1405 |               "      <th>Average</th>\n",
1406 |               "      <th>Balls_Faced</th>\n",
1407 |               "      <th>Strike_Rate</th>\n",
1408 |               "      <th>Centuries</th>\n",
1409 |               "      <th>Half_centuries</th>\n",
1410 |               "      <th>Ducks</th>\n",
1411 |               "      <th>Fours</th>\n",
1412 |               "      <th>Sixes</th>\n",
1413 |               "      <th>Player_Name</th>\n",
1414 |               "      <th>Country</th>\n",
1415 |               "    </tr>\n",
1416 |               "  </thead>\n",
1417 |               "  <tbody>\n",
1418 |               "    <tr>\n",
1419 |               "      <th>0</th>\n",
1420 |               "      <td>1989-2012</td>\n",
1421 |               "      <td>463</td>\n",
1422 |               "      <td>452</td>\n",
1423 |               "      <td>41</td>\n",
1424 |               "      <td>18426</td>\n",
1425 |               "      <td>200*</td>\n",
1426 |               "      <td>44.83</td>\n",
1427 |               "      <td>21368</td>\n",
1428 |               "      <td>86.23</td>\n",
1429 |               "      <td>49</td>\n",
1430 |               "      <td>96</td>\n",
1431 |               "      <td>20</td>\n",
1432 |               "      <td>2016</td>\n",
1433 |               "      <td>195</td>\n",
1434 |               "      <td>SR Tendulkar</td>\n",
1435 |               "      <td>INDIA)</td>\n",
1436 |               "    </tr>\n",
1437 |               "    <tr>\n",
1438 |               "      <th>1</th>\n",
1439 |               "      <td>2000-2015</td>\n",
1440 |               "      <td>404</td>\n",
1441 |               "      <td>380</td>\n",
1442 |               "      <td>41</td>\n",
1443 |               "      <td>14234</td>\n",
1444 |               "      <td>169</td>\n",
1445 |               "      <td>41.98</td>\n",
1446 |               "      <td>18048</td>\n",
1447 |               "      <td>78.86</td>\n",
1448 |               "      <td>25</td>\n",
1449 |               "      <td>93</td>\n",
1450 |               "      <td>15</td>\n",
1451 |               "      <td>1385</td>\n",
1452 |               "      <td>88</td>\n",
1453 |               "      <td>KC Sangakkara</td>\n",
1454 |               "      <td>Asia/ICC/SL)</td>\n",
1455 |               "    </tr>\n",
1456 |               "    <tr>\n",
1457 |               "      <th>2</th>\n",
1458 |               "      <td>1995-2012</td>\n",
1459 |               "      <td>375</td>\n",
1460 |               "      <td>365</td>\n",
1461 |               "      <td>39</td>\n",
1462 |               "      <td>13704</td>\n",
1463 |               "      <td>164</td>\n",
1464 |               "      <td>42.03</td>\n",
1465 |               "      <td>17046</td>\n",
1466 |               "      <td>80.39</td>\n",
1467 |               "      <td>30</td>\n",
1468 |               "      <td>82</td>\n",
1469 |               "      <td>20</td>\n",
1470 |               "      <td>1231</td>\n",
1471 |               "      <td>162</td>\n",
1472 |               "      <td>RT Ponting</td>\n",
1473 |               "      <td>AUS/ICC)</td>\n",
1474 |               "    </tr>\n",
1475 |               "    <tr>\n",
1476 |               "      <th>3</th>\n",
1477 |               "      <td>1989-2011</td>\n",
1478 |               "      <td>445</td>\n",
1479 |               "      <td>433</td>\n",
1480 |               "      <td>18</td>\n",
1481 |               "      <td>13430</td>\n",
1482 |               "      <td>189</td>\n",
1483 |               "      <td>32.36</td>\n",
1484 |               "      <td>14725</td>\n",
1485 |               "      <td>91.20</td>\n",
1486 |               "      <td>28</td>\n",
1487 |               "      <td>68</td>\n",
1488 |               "      <td>34</td>\n",
1489 |               "      <td>1500</td>\n",
1490 |               "      <td>270</td>\n",
1491 |               "      <td>ST Jayasuriya</td>\n",
1492 |               "      <td>Asia/SL)</td>\n",
1493 |               "    </tr>\n",
1494 |               "    <tr>\n",
1495 |               "      <th>4</th>\n",
1496 |               "      <td>1998-2015</td>\n",
1497 |               "      <td>448</td>\n",
1498 |               "      <td>418</td>\n",
1499 |               "      <td>39</td>\n",
1500 |               "      <td>12650</td>\n",
1501 |               "      <td>144</td>\n",
1502 |               "      <td>33.37</td>\n",
1503 |               "      <td>16020</td>\n",
1504 |               "      <td>78.96</td>\n",
1505 |               "      <td>19</td>\n",
1506 |               "      <td>77</td>\n",
1507 |               "      <td>28</td>\n",
1508 |               "      <td>1119</td>\n",
1509 |               "      <td>76</td>\n",
1510 |               "      <td>DPMD Jayawardene</td>\n",
1511 |               "      <td>Asia/SL)</td>\n",
1512 |               "    </tr>\n",
1513 |               "  </tbody>\n",
1514 |               "</table>\n",
1515 |               "</div>"
1516 |             ],
1517 |             "text/plain": [
1518 |               "        Span  Match  Innings  NotOut   Runs Highest_score  Average  \\\n",
1519 |               "0  1989-2012    463      452      41  18426          200*    44.83   \n",
1520 |               "1  2000-2015    404      380      41  14234           169    41.98   \n",
1521 |               "2  1995-2012    375      365      39  13704           164    42.03   \n",
1522 |               "3  1989-2011    445      433      18  13430           189    32.36   \n",
1523 |               "4  1998-2015    448      418      39  12650           144    33.37   \n",
1524 |               "\n",
1525 |               "   Balls_Faced  Strike_Rate  Centuries  Half_centuries  Ducks Fours Sixes  \\\n",
1526 |               "0        21368        86.23         49              96     20  2016   195   \n",
1527 |               "1        18048        78.86         25              93     15  1385    88   \n",
1528 |               "2        17046        80.39         30              82     20  1231   162   \n",
1529 |               "3        14725        91.20         28              68     34  1500   270   \n",
1530 |               "4        16020        78.96         19              77     28  1119    76   \n",
1531 |               "\n",
1532 |               "         Player_Name       Country  \n",
1533 |               "0      SR Tendulkar         INDIA)  \n",
1534 |               "1     KC Sangakkara   Asia/ICC/SL)  \n",
1535 |               "2        RT Ponting       AUS/ICC)  \n",
1536 |               "3     ST Jayasuriya       Asia/SL)  \n",
1537 |               "4  DPMD Jayawardene       Asia/SL)  "
1538 |             ]
1539 |           },
1540 |           "metadata": {},
1541 |           "output_type": "display_data"
1542 |         }
1543 |       ],
1544 |       "source": [
1545 |         "# line 1\n",
1546 |         "# df = df.drop('Player', axis=1)\n",
1547 |         "\n",
1548 |         "# line 2\n",
1549 |         "df.drop('Player', axis=1, inplace=True)\n",
1550 |         "\n",
1551 |         "# line 1 and line 2 both are same\n",
1552 |         "\n",
1553 |         "display(df.head())"
1554 |       ]
1555 |     },
1556 |     {
1557 |       "cell_type": "markdown",
1558 |       "metadata": {
1559 |         "id": "CcpvRLtEv8hs"
1560 |       },
1561 |       "source": [
1562 |         "#### How to replace/remove a value from a pandas column?"
1563 |       ]
1564 |     },
1565 |     {
1566 |       "cell_type": "code",
1567 |       "execution_count": null,
1568 |       "metadata": {
1569 |         "id": "bKMHBsQEv8hs",
1570 |         "outputId": "478310ad-f494-4e29-b87c-1fd31714c510"
1571 |       },
1572 |       "outputs": [
1573 |         {
1574 |           "data": {
1575 |             "text/html": [
1576 |               "<div>\n",
1577 |               "<style scoped>\n",
1578 |               "    .dataframe tbody tr th:only-of-type {\n",
1579 |               "        vertical-align: middle;\n",
1580 |               "    }\n",
1581 |               "\n",
1582 |               "    .dataframe tbody tr th {\n",
1583 |               "        vertical-align: top;\n",
1584 |               "    }\n",
1585 |               "\n",
1586 |               "    .dataframe thead th {\n",
1587 |               "        text-align: right;\n",
1588 |               "    }\n",
1589 |               "</style>\n",
1590 |               "<table border=\"1\" class=\"dataframe\">\n",
1591 |               "  <thead>\n",
1592 |               "    <tr style=\"text-align: right;\">\n",
1593 |               "      <th></th>\n",
1594 |               "      <th>Span</th>\n",
1595 |               "      <th>Match</th>\n",
1596 |               "      <th>Innings</th>\n",
1597 |               "      <th>NotOut</th>\n",
1598 |               "      <th>Runs</th>\n",
1599 |               "      <th>Highest_score</th>\n",
1600 |               "      <th>Average</th>\n",
1601 |               "      <th>Balls_Faced</th>\n",
1602 |               "      <th>Strike_Rate</th>\n",
1603 |               "      <th>Centuries</th>\n",
1604 |               "      <th>Half_centuries</th>\n",
1605 |               "      <th>Ducks</th>\n",
1606 |               "      <th>Fours</th>\n",
1607 |               "      <th>Sixes</th>\n",
1608 |               "      <th>Player_Name</th>\n",
1609 |               "      <th>Country</th>\n",
1610 |               "    </tr>\n",
1611 |               "  </thead>\n",
1612 |               "  <tbody>\n",
1613 |               "    <tr>\n",
1614 |               "      <th>0</th>\n",
1615 |               "      <td>1989-2012</td>\n",
1616 |               "      <td>463</td>\n",
1617 |               "      <td>452</td>\n",
1618 |               "      <td>41</td>\n",
1619 |               "      <td>18426</td>\n",
1620 |               "      <td>200*</td>\n",
1621 |               "      <td>44.83</td>\n",
1622 |               "      <td>21368</td>\n",
1623 |               "      <td>86.23</td>\n",
1624 |               "      <td>49</td>\n",
1625 |               "      <td>96</td>\n",
1626 |               "      <td>20</td>\n",
1627 |               "      <td>2016</td>\n",
1628 |               "      <td>195</td>\n",
1629 |               "      <td>SR Tendulkar</td>\n",
1630 |               "      <td>INDIA</td>\n",
1631 |               "    </tr>\n",
1632 |               "    <tr>\n",
1633 |               "      <th>1</th>\n",
1634 |               "      <td>2000-2015</td>\n",
1635 |               "      <td>404</td>\n",
1636 |               "      <td>380</td>\n",
1637 |               "      <td>41</td>\n",
1638 |               "      <td>14234</td>\n",
1639 |               "      <td>169</td>\n",
1640 |               "      <td>41.98</td>\n",
1641 |               "      <td>18048</td>\n",
1642 |               "      <td>78.86</td>\n",
1643 |               "      <td>25</td>\n",
1644 |               "      <td>93</td>\n",
1645 |               "      <td>15</td>\n",
1646 |               "      <td>1385</td>\n",
1647 |               "      <td>88</td>\n",
1648 |               "      <td>KC Sangakkara</td>\n",
1649 |               "      <td>Asia/ICC/SL</td>\n",
1650 |               "    </tr>\n",
1651 |               "    <tr>\n",
1652 |               "      <th>2</th>\n",
1653 |               "      <td>1995-2012</td>\n",
1654 |               "      <td>375</td>\n",
1655 |               "      <td>365</td>\n",
1656 |               "      <td>39</td>\n",
1657 |               "      <td>13704</td>\n",
1658 |               "      <td>164</td>\n",
1659 |               "      <td>42.03</td>\n",
1660 |               "      <td>17046</td>\n",
1661 |               "      <td>80.39</td>\n",
1662 |               "      <td>30</td>\n",
1663 |               "      <td>82</td>\n",
1664 |               "      <td>20</td>\n",
1665 |               "      <td>1231</td>\n",
1666 |               "      <td>162</td>\n",
1667 |               "      <td>RT Ponting</td>\n",
1668 |               "      <td>AUS/ICC</td>\n",
1669 |               "    </tr>\n",
1670 |               "    <tr>\n",
1671 |               "      <th>3</th>\n",
1672 |               "      <td>1989-2011</td>\n",
1673 |               "      <td>445</td>\n",
1674 |               "      <td>433</td>\n",
1675 |               "      <td>18</td>\n",
1676 |               "      <td>13430</td>\n",
1677 |               "      <td>189</td>\n",
1678 |               "      <td>32.36</td>\n",
1679 |               "      <td>14725</td>\n",
1680 |               "      <td>91.20</td>\n",
1681 |               "      <td>28</td>\n",
1682 |               "      <td>68</td>\n",
1683 |               "      <td>34</td>\n",
1684 |               "      <td>1500</td>\n",
1685 |               "      <td>270</td>\n",
1686 |               "      <td>ST Jayasuriya</td>\n",
1687 |               "      <td>Asia/SL</td>\n",
1688 |               "    </tr>\n",
1689 |               "    <tr>\n",
1690 |               "      <th>4</th>\n",
1691 |               "      <td>1998-2015</td>\n",
1692 |               "      <td>448</td>\n",
1693 |               "      <td>418</td>\n",
1694 |               "      <td>39</td>\n",
1695 |               "      <td>12650</td>\n",
1696 |               "      <td>144</td>\n",
1697 |               "      <td>33.37</td>\n",
1698 |               "      <td>16020</td>\n",
1699 |               "      <td>78.96</td>\n",
1700 |               "      <td>19</td>\n",
1701 |               "      <td>77</td>\n",
1702 |               "      <td>28</td>\n",
1703 |               "      <td>1119</td>\n",
1704 |               "      <td>76</td>\n",
1705 |               "      <td>DPMD Jayawardene</td>\n",
1706 |               "      <td>Asia/SL</td>\n",
1707 |               "    </tr>\n",
1708 |               "  </tbody>\n",
1709 |               "</table>\n",
1710 |               "</div>"
1711 |             ],
1712 |             "text/plain": [
1713 |               "        Span  Match  Innings  NotOut   Runs Highest_score  Average  \\\n",
1714 |               "0  1989-2012    463      452      41  18426          200*    44.83   \n",
1715 |               "1  2000-2015    404      380      41  14234           169    41.98   \n",
1716 |               "2  1995-2012    375      365      39  13704           164    42.03   \n",
1717 |               "3  1989-2011    445      433      18  13430           189    32.36   \n",
1718 |               "4  1998-2015    448      418      39  12650           144    33.37   \n",
1719 |               "\n",
1720 |               "   Balls_Faced  Strike_Rate  Centuries  Half_centuries  Ducks Fours Sixes  \\\n",
1721 |               "0        21368        86.23         49              96     20  2016   195   \n",
1722 |               "1        18048        78.86         25              93     15  1385    88   \n",
1723 |               "2        17046        80.39         30              82     20  1231   162   \n",
1724 |               "3        14725        91.20         28              68     34  1500   270   \n",
1725 |               "4        16020        78.96         19              77     28  1119    76   \n",
1726 |               "\n",
1727 |               "         Player_Name      Country  \n",
1728 |               "0      SR Tendulkar         INDIA  \n",
1729 |               "1     KC Sangakkara   Asia/ICC/SL  \n",
1730 |               "2        RT Ponting       AUS/ICC  \n",
1731 |               "3     ST Jayasuriya       Asia/SL  \n",
1732 |               "4  DPMD Jayawardene       Asia/SL  "
1733 |             ]
1734 |           },
1735 |           "metadata": {},
1736 |           "output_type": "display_data"
1737 |         }
1738 |       ],
1739 |       "source": [
1740 |         "df['Country'] = df['Country'].str.replace(\")\", \"\")\n",
1741 |         "\n",
1742 |         "display(df.head())"
1743 |       ]
1744 |     },
1745 |     {
1746 |       "cell_type": "code",
1747 |       "execution_count": null,
1748 |       "metadata": {
1749 |         "id": "z6o5XnAcv8ht",
1750 |         "outputId": "63a9a402-1f01-4158-d70c-0a84ed96ee05"
1751 |       },
1752 |       "outputs": [
1753 |         {
1754 |           "name": "stdout",
1755 |           "output_type": "stream",
1756 |           "text": [
1757 |             "Index(['Span', 'Match', 'Innings', 'NotOut', 'Runs', 'Highest_score',\n",
1758 |             "       'Average', 'Balls_Faced', 'Strike_Rate', 'Centuries', 'Half_centuries',\n",
1759 |             "       'Ducks', 'Fours', 'Sixes', 'Player_Name', 'Country'],\n",
1760 |             "      dtype='object')\n"
1761 |           ]
1762 |         }
1763 |       ],
1764 |       "source": [
1765 |         "print(df.columns)\n",
1766 |         "\n",
1767 |         "new_col_sequence = ['Player_Name', 'Country', 'Span', 'Match', 'Innings', 'NotOut', 'Runs', 'Highest_score',\n",
1768 |         "       'Average', 'Balls_Faced', 'Strike_Rate', 'Centuries', 'Half_centuries', 'Ducks', 'Fours', 'Sixes']"
1769 |       ]
1770 |     },
1771 |     {
1772 |       "cell_type": "code",
1773 |       "execution_count": null,
1774 |       "metadata": {
1775 |         "id": "afGc2B3Rv8ht",
1776 |         "outputId": "153f69a2-46d8-4eaa-fde7-6bd4f41fd365"
1777 |       },
1778 |       "outputs": [
1779 |         {
1780 |           "data": {
1781 |             "text/html": [
1782 |               "<div>\n",
1783 |               "<style scoped>\n",
1784 |               "    .dataframe tbody tr th:only-of-type {\n",
1785 |               "        vertical-align: middle;\n",
1786 |               "    }\n",
1787 |               "\n",
1788 |               "    .dataframe tbody tr th {\n",
1789 |               "        vertical-align: top;\n",
1790 |               "    }\n",
1791 |               "\n",
1792 |               "    .dataframe thead th {\n",
1793 |               "        text-align: right;\n",
1794 |               "    }\n",
1795 |               "</style>\n",
1796 |               "<table border=\"1\" class=\"dataframe\">\n",
1797 |               "  <thead>\n",
1798 |               "    <tr style=\"text-align: right;\">\n",
1799 |               "      <th></th>\n",
1800 |               "      <th>Player_Name</th>\n",
1801 |               "      <th>Country</th>\n",
1802 |               "      <th>Span</th>\n",
1803 |               "      <th>Match</th>\n",
1804 |               "      <th>Innings</th>\n",
1805 |               "      <th>NotOut</th>\n",
1806 |               "      <th>Runs</th>\n",
1807 |               "      <th>Highest_score</th>\n",
1808 |               "      <th>Average</th>\n",
1809 |               "      <th>Balls_Faced</th>\n",
1810 |               "      <th>Strike_Rate</th>\n",
1811 |               "      <th>Centuries</th>\n",
1812 |               "      <th>Half_centuries</th>\n",
1813 |               "      <th>Ducks</th>\n",
1814 |               "      <th>Fours</th>\n",
1815 |               "      <th>Sixes</th>\n",
1816 |               "    </tr>\n",
1817 |               "  </thead>\n",
1818 |               "  <tbody>\n",
1819 |               "    <tr>\n",
1820 |               "      <th>0</th>\n",
1821 |               "      <td>SR Tendulkar</td>\n",
1822 |               "      <td>INDIA</td>\n",
1823 |               "      <td>1989-2012</td>\n",
1824 |               "      <td>463</td>\n",
1825 |               "      <td>452</td>\n",
1826 |               "      <td>41</td>\n",
1827 |               "      <td>18426</td>\n",
1828 |               "      <td>200*</td>\n",
1829 |               "      <td>44.83</td>\n",
1830 |               "      <td>21368</td>\n",
1831 |               "      <td>86.23</td>\n",
1832 |               "      <td>49</td>\n",
1833 |               "      <td>96</td>\n",
1834 |               "      <td>20</td>\n",
1835 |               "      <td>2016</td>\n",
1836 |               "      <td>195</td>\n",
1837 |               "    </tr>\n",
1838 |               "    <tr>\n",
1839 |               "      <th>1</th>\n",
1840 |               "      <td>KC Sangakkara</td>\n",
1841 |               "      <td>Asia/ICC/SL</td>\n",
1842 |               "      <td>2000-2015</td>\n",
1843 |               "      <td>404</td>\n",
1844 |               "      <td>380</td>\n",
1845 |               "      <td>41</td>\n",
1846 |               "      <td>14234</td>\n",
1847 |               "      <td>169</td>\n",
1848 |               "      <td>41.98</td>\n",
1849 |               "      <td>18048</td>\n",
1850 |               "      <td>78.86</td>\n",
1851 |               "      <td>25</td>\n",
1852 |               "      <td>93</td>\n",
1853 |               "      <td>15</td>\n",
1854 |               "      <td>1385</td>\n",
1855 |               "      <td>88</td>\n",
1856 |               "    </tr>\n",
1857 |               "    <tr>\n",
1858 |               "      <th>2</th>\n",
1859 |               "      <td>RT Ponting</td>\n",
1860 |               "      <td>AUS/ICC</td>\n",
1861 |               "      <td>1995-2012</td>\n",
1862 |               "      <td>375</td>\n",
1863 |               "      <td>365</td>\n",
1864 |               "      <td>39</td>\n",
1865 |               "      <td>13704</td>\n",
1866 |               "      <td>164</td>\n",
1867 |               "      <td>42.03</td>\n",
1868 |               "      <td>17046</td>\n",
1869 |               "      <td>80.39</td>\n",
1870 |               "      <td>30</td>\n",
1871 |               "      <td>82</td>\n",
1872 |               "      <td>20</td>\n",
1873 |               "      <td>1231</td>\n",
1874 |               "      <td>162</td>\n",
1875 |               "    </tr>\n",
1876 |               "    <tr>\n",
1877 |               "      <th>3</th>\n",
1878 |               "      <td>ST Jayasuriya</td>\n",
1879 |               "      <td>Asia/SL</td>\n",
1880 |               "      <td>1989-2011</td>\n",
1881 |               "      <td>445</td>\n",
1882 |               "      <td>433</td>\n",
1883 |               "      <td>18</td>\n",
1884 |               "      <td>13430</td>\n",
1885 |               "      <td>189</td>\n",
1886 |               "      <td>32.36</td>\n",
1887 |               "      <td>14725</td>\n",
1888 |               "      <td>91.20</td>\n",
1889 |               "      <td>28</td>\n",
1890 |               "      <td>68</td>\n",
1891 |               "      <td>34</td>\n",
1892 |               "      <td>1500</td>\n",
1893 |               "      <td>270</td>\n",
1894 |               "    </tr>\n",
1895 |               "    <tr>\n",
1896 |               "      <th>4</th>\n",
1897 |               "      <td>DPMD Jayawardene</td>\n",
1898 |               "      <td>Asia/SL</td>\n",
1899 |               "      <td>1998-2015</td>\n",
1900 |               "      <td>448</td>\n",
1901 |               "      <td>418</td>\n",
1902 |               "      <td>39</td>\n",
1903 |               "      <td>12650</td>\n",
1904 |               "      <td>144</td>\n",
1905 |               "      <td>33.37</td>\n",
1906 |               "      <td>16020</td>\n",
1907 |               "      <td>78.96</td>\n",
1908 |               "      <td>19</td>\n",
1909 |               "      <td>77</td>\n",
1910 |               "      <td>28</td>\n",
1911 |               "      <td>1119</td>\n",
1912 |               "      <td>76</td>\n",
1913 |               "    </tr>\n",
1914 |               "  </tbody>\n",
1915 |               "</table>\n",
1916 |               "</div>"
1917 |             ],
1918 |             "text/plain": [
1919 |               "         Player_Name      Country       Span  Match  Innings  NotOut   Runs  \\\n",
1920 |               "0      SR Tendulkar         INDIA  1989-2012    463      452      41  18426   \n",
1921 |               "1     KC Sangakkara   Asia/ICC/SL  2000-2015    404      380      41  14234   \n",
1922 |               "2        RT Ponting       AUS/ICC  1995-2012    375      365      39  13704   \n",
1923 |               "3     ST Jayasuriya       Asia/SL  1989-2011    445      433      18  13430   \n",
1924 |               "4  DPMD Jayawardene       Asia/SL  1998-2015    448      418      39  12650   \n",
1925 |               "\n",
1926 |               "  Highest_score  Average  Balls_Faced  Strike_Rate  Centuries  Half_centuries  \\\n",
1927 |               "0          200*    44.83        21368        86.23         49              96   \n",
1928 |               "1           169    41.98        18048        78.86         25              93   \n",
1929 |               "2           164    42.03        17046        80.39         30              82   \n",
1930 |               "3           189    32.36        14725        91.20         28              68   \n",
1931 |               "4           144    33.37        16020        78.96         19              77   \n",
1932 |               "\n",
1933 |               "   Ducks Fours Sixes  \n",
1934 |               "0     20  2016   195  \n",
1935 |               "1     15  1385    88  \n",
1936 |               "2     20  1231   162  \n",
1937 |               "3     34  1500   270  \n",
1938 |               "4     28  1119    76  "
1939 |             ]
1940 |           },
1941 |           "metadata": {},
1942 |           "output_type": "display_data"
1943 |         }
1944 |       ],
1945 |       "source": [
1946 |         "df = df[new_col_sequence]\n",
1947 |         "\n",
1948 |         "display(df.head())"
1949 |       ]
1950 |     },
1951 |     {
1952 |       "cell_type": "code",
1953 |       "execution_count": null,
1954 |       "metadata": {
1955 |         "id": "Nlyq488Zv8ht"
1956 |       },
1957 |       "outputs": [],
1958 |       "source": [
1959 |         ""
1960 |       ]
1961 |     }
1962 |   ],
1963 |   "metadata": {
1964 |     "colab": {
1965 |       "name": "Lecture_02.ipynb",
1966 |       "provenance": [],
1967 |       "include_colab_link": true
1968 |     },
1969 |     "kernelspec": {
1970 |       "display_name": "Python 3",
1971 |       "language": "python",
1972 |       "name": "python3"
1973 |     },
1974 |     "language_info": {
1975 |       "codemirror_mode": {
1976 |         "name": "ipython",
1977 |         "version": 3
1978 |       },
1979 |       "file_extension": ".py",
1980 |       "mimetype": "text/x-python",
1981 |       "name": "python",
1982 |       "nbconvert_exporter": "python",
1983 |       "pygments_lexer": "ipython3",
1984 |       "version": "3.7.7"
1985 |     }
1986 |   },
1987 |   "nbformat": 4,
1988 |   "nbformat_minor": 0
1989 | }
1990 | 


--------------------------------------------------------------------------------
/8. Categorical Variable Encoding.ipynb:
--------------------------------------------------------------------------------
   1 | {
   2 |  "cells": [
   3 |   {
   4 |    "cell_type": "markdown",
   5 |    "metadata": {},
   6 |    "source": [
   7 |     "# Lecture 8: Categorical Variable Encoding\n",
   8 |     "\n",
   9 |     "Instructor: Md Shahidullah Kawsar\n",
  10 |     "<br>Data Scientist, IDARE, Houston, TX, USA\n",
  11 |     "\n",
  12 |     "#### Objectives:\n",
  13 |     "- Dealing with categorical variables\n",
  14 |     "- Label encoding\n",
  15 |     "- One-hot encoding\n",
  16 |     "- Categorical variable creation from the numeric variable\n",
  17 |     "\n",
  18 |     "#### References:\n",
  19 |     "<br>[1] One-Hot Encoding vs. Label Encoding using Scikit-Learn: https://www.analyticsvidhya.com/blog/2020/03/one-hot-encoding-vs-label-encoding-using-scikit-learn/\n",
  20 |     "<br>[2] Label Encoding: https://scikit-learn.org/stable/modules/generated/sklearn.preprocessing.LabelEncoder.html\n",
  21 |     "<br>[3] One-hot encoding: https://pandas.pydata.org/docs/reference/api/pandas.get_dummies.html\n",
  22 |     "<br>[4] https://pandas.pydata.org/docs/reference/api/pandas.cut.html"
  23 |    ]
  24 |   },
  25 |   {
  26 |    "cell_type": "code",
  27 |    "execution_count": 51,
  28 |    "metadata": {},
  29 |    "outputs": [],
  30 |    "source": [
  31 |     "import pandas as pd\n",
  32 |     "\n",
  33 |     "import seaborn as sns\n",
  34 |     "import matplotlib.pyplot as plt\n",
  35 |     "\n",
  36 |     "from sklearn.preprocessing import LabelEncoder\n",
  37 |     "\n",
  38 |     "sns.set_context(\"talk\")"
  39 |    ]
  40 |   },
  41 |   {
  42 |    "cell_type": "markdown",
  43 |    "metadata": {},
  44 |    "source": [
  45 |     "#### Load data"
  46 |    ]
  47 |   },
  48 |   {
  49 |    "cell_type": "code",
  50 |    "execution_count": 52,
  51 |    "metadata": {},
  52 |    "outputs": [
  53 |     {
  54 |      "data": {
  55 |       "text/html": [
  56 |        "<div>\n",
  57 |        "<style scoped>\n",
  58 |        "    .dataframe tbody tr th:only-of-type {\n",
  59 |        "        vertical-align: middle;\n",
  60 |        "    }\n",
  61 |        "\n",
  62 |        "    .dataframe tbody tr th {\n",
  63 |        "        vertical-align: top;\n",
  64 |        "    }\n",
  65 |        "\n",
  66 |        "    .dataframe thead th {\n",
  67 |        "        text-align: right;\n",
  68 |        "    }\n",
  69 |        "</style>\n",
  70 |        "<table border=\"1\" class=\"dataframe\">\n",
  71 |        "  <thead>\n",
  72 |        "    <tr style=\"text-align: right;\">\n",
  73 |        "      <th></th>\n",
  74 |        "      <th>model</th>\n",
  75 |        "      <th>year</th>\n",
  76 |        "      <th>price</th>\n",
  77 |        "      <th>transmission</th>\n",
  78 |        "      <th>mileage</th>\n",
  79 |        "      <th>fuelType</th>\n",
  80 |        "      <th>tax</th>\n",
  81 |        "      <th>mpg</th>\n",
  82 |        "      <th>engineSize</th>\n",
  83 |        "    </tr>\n",
  84 |        "  </thead>\n",
  85 |        "  <tbody>\n",
  86 |        "    <tr>\n",
  87 |        "      <th>0</th>\n",
  88 |        "      <td>5 Series</td>\n",
  89 |        "      <td>2014</td>\n",
  90 |        "      <td>11200</td>\n",
  91 |        "      <td>Automatic</td>\n",
  92 |        "      <td>67068</td>\n",
  93 |        "      <td>Diesel</td>\n",
  94 |        "      <td>125</td>\n",
  95 |        "      <td>57.6</td>\n",
  96 |        "      <td>2.0</td>\n",
  97 |        "    </tr>\n",
  98 |        "    <tr>\n",
  99 |        "      <th>1</th>\n",
 100 |        "      <td>6 Series</td>\n",
 101 |        "      <td>2018</td>\n",
 102 |        "      <td>27000</td>\n",
 103 |        "      <td>Automatic</td>\n",
 104 |        "      <td>14827</td>\n",
 105 |        "      <td>Petrol</td>\n",
 106 |        "      <td>145</td>\n",
 107 |        "      <td>42.8</td>\n",
 108 |        "      <td>2.0</td>\n",
 109 |        "    </tr>\n",
 110 |        "    <tr>\n",
 111 |        "      <th>2</th>\n",
 112 |        "      <td>5 Series</td>\n",
 113 |        "      <td>2016</td>\n",
 114 |        "      <td>16000</td>\n",
 115 |        "      <td>Automatic</td>\n",
 116 |        "      <td>62794</td>\n",
 117 |        "      <td>Diesel</td>\n",
 118 |        "      <td>160</td>\n",
 119 |        "      <td>51.4</td>\n",
 120 |        "      <td>3.0</td>\n",
 121 |        "    </tr>\n",
 122 |        "    <tr>\n",
 123 |        "      <th>3</th>\n",
 124 |        "      <td>1 Series</td>\n",
 125 |        "      <td>2017</td>\n",
 126 |        "      <td>12750</td>\n",
 127 |        "      <td>Automatic</td>\n",
 128 |        "      <td>26676</td>\n",
 129 |        "      <td>Diesel</td>\n",
 130 |        "      <td>145</td>\n",
 131 |        "      <td>72.4</td>\n",
 132 |        "      <td>1.5</td>\n",
 133 |        "    </tr>\n",
 134 |        "    <tr>\n",
 135 |        "      <th>4</th>\n",
 136 |        "      <td>7 Series</td>\n",
 137 |        "      <td>2014</td>\n",
 138 |        "      <td>14500</td>\n",
 139 |        "      <td>Automatic</td>\n",
 140 |        "      <td>39554</td>\n",
 141 |        "      <td>Diesel</td>\n",
 142 |        "      <td>160</td>\n",
 143 |        "      <td>50.4</td>\n",
 144 |        "      <td>3.0</td>\n",
 145 |        "    </tr>\n",
 146 |        "    <tr>\n",
 147 |        "      <th>5</th>\n",
 148 |        "      <td>5 Series</td>\n",
 149 |        "      <td>2016</td>\n",
 150 |        "      <td>14900</td>\n",
 151 |        "      <td>Automatic</td>\n",
 152 |        "      <td>35309</td>\n",
 153 |        "      <td>Diesel</td>\n",
 154 |        "      <td>125</td>\n",
 155 |        "      <td>60.1</td>\n",
 156 |        "      <td>2.0</td>\n",
 157 |        "    </tr>\n",
 158 |        "    <tr>\n",
 159 |        "      <th>6</th>\n",
 160 |        "      <td>5 Series</td>\n",
 161 |        "      <td>2017</td>\n",
 162 |        "      <td>16000</td>\n",
 163 |        "      <td>Automatic</td>\n",
 164 |        "      <td>38538</td>\n",
 165 |        "      <td>Diesel</td>\n",
 166 |        "      <td>125</td>\n",
 167 |        "      <td>60.1</td>\n",
 168 |        "      <td>2.0</td>\n",
 169 |        "    </tr>\n",
 170 |        "    <tr>\n",
 171 |        "      <th>7</th>\n",
 172 |        "      <td>2 Series</td>\n",
 173 |        "      <td>2018</td>\n",
 174 |        "      <td>16250</td>\n",
 175 |        "      <td>Manual</td>\n",
 176 |        "      <td>10401</td>\n",
 177 |        "      <td>Petrol</td>\n",
 178 |        "      <td>145</td>\n",
 179 |        "      <td>52.3</td>\n",
 180 |        "      <td>1.5</td>\n",
 181 |        "    </tr>\n",
 182 |        "    <tr>\n",
 183 |        "      <th>8</th>\n",
 184 |        "      <td>4 Series</td>\n",
 185 |        "      <td>2017</td>\n",
 186 |        "      <td>14250</td>\n",
 187 |        "      <td>Manual</td>\n",
 188 |        "      <td>42668</td>\n",
 189 |        "      <td>Diesel</td>\n",
 190 |        "      <td>30</td>\n",
 191 |        "      <td>62.8</td>\n",
 192 |        "      <td>2.0</td>\n",
 193 |        "    </tr>\n",
 194 |        "    <tr>\n",
 195 |        "      <th>9</th>\n",
 196 |        "      <td>5 Series</td>\n",
 197 |        "      <td>2016</td>\n",
 198 |        "      <td>14250</td>\n",
 199 |        "      <td>Automatic</td>\n",
 200 |        "      <td>36099</td>\n",
 201 |        "      <td>Diesel</td>\n",
 202 |        "      <td>20</td>\n",
 203 |        "      <td>68.9</td>\n",
 204 |        "      <td>2.0</td>\n",
 205 |        "    </tr>\n",
 206 |        "  </tbody>\n",
 207 |        "</table>\n",
 208 |        "</div>"
 209 |       ],
 210 |       "text/plain": [
 211 |        "       model  year  price transmission  mileage fuelType  tax   mpg  \\\n",
 212 |        "0   5 Series  2014  11200    Automatic    67068   Diesel  125  57.6   \n",
 213 |        "1   6 Series  2018  27000    Automatic    14827   Petrol  145  42.8   \n",
 214 |        "2   5 Series  2016  16000    Automatic    62794   Diesel  160  51.4   \n",
 215 |        "3   1 Series  2017  12750    Automatic    26676   Diesel  145  72.4   \n",
 216 |        "4   7 Series  2014  14500    Automatic    39554   Diesel  160  50.4   \n",
 217 |        "5   5 Series  2016  14900    Automatic    35309   Diesel  125  60.1   \n",
 218 |        "6   5 Series  2017  16000    Automatic    38538   Diesel  125  60.1   \n",
 219 |        "7   2 Series  2018  16250       Manual    10401   Petrol  145  52.3   \n",
 220 |        "8   4 Series  2017  14250       Manual    42668   Diesel   30  62.8   \n",
 221 |        "9   5 Series  2016  14250    Automatic    36099   Diesel   20  68.9   \n",
 222 |        "\n",
 223 |        "   engineSize  \n",
 224 |        "0         2.0  \n",
 225 |        "1         2.0  \n",
 226 |        "2         3.0  \n",
 227 |        "3         1.5  \n",
 228 |        "4         3.0  \n",
 229 |        "5         2.0  \n",
 230 |        "6         2.0  \n",
 231 |        "7         1.5  \n",
 232 |        "8         2.0  \n",
 233 |        "9         2.0  "
 234 |       ]
 235 |      },
 236 |      "metadata": {},
 237 |      "output_type": "display_data"
 238 |     }
 239 |    ],
 240 |    "source": [
 241 |     "df = pd.read_csv(\"bmw.csv\")\n",
 242 |     "\n",
 243 |     "display(df.head(10))"
 244 |    ]
 245 |   },
 246 |   {
 247 |    "cell_type": "markdown",
 248 |    "metadata": {},
 249 |    "source": [
 250 |     "#### Dealing with categorical variables"
 251 |    ]
 252 |   },
 253 |   {
 254 |    "cell_type": "code",
 255 |    "execution_count": 53,
 256 |    "metadata": {},
 257 |    "outputs": [
 258 |     {
 259 |      "name": "stdout",
 260 |      "output_type": "stream",
 261 |      "text": [
 262 |       "[' 5 Series' ' 6 Series' ' 1 Series' ' 7 Series' ' 2 Series' ' 4 Series'\n",
 263 |       " ' X3' ' 3 Series' ' X5' ' X4' ' i3' ' X1' ' M4' ' X2' ' X6' ' 8 Series'\n",
 264 |       " ' Z4' ' X7' ' M5' ' i8' ' M2' ' M3' ' M6' ' Z3']\n",
 265 |       "24\n",
 266 |       "['Automatic' 'Manual' 'Semi-Auto']\n",
 267 |       "3\n",
 268 |       "['Diesel' 'Petrol' 'Other' 'Hybrid' 'Electric']\n",
 269 |       "5\n"
 270 |      ]
 271 |     }
 272 |    ],
 273 |    "source": [
 274 |     "print(df[\"model\"].unique())\n",
 275 |     "print(len(df[\"model\"].unique()))\n",
 276 |     "\n",
 277 |     "print(df[\"transmission\"].unique())\n",
 278 |     "print(len(df[\"transmission\"].unique()))\n",
 279 |     "\n",
 280 |     "print(df[\"fuelType\"].unique())\n",
 281 |     "print(len(df[\"fuelType\"].unique()))"
 282 |    ]
 283 |   },
 284 |   {
 285 |    "cell_type": "markdown",
 286 |    "metadata": {},
 287 |    "source": [
 288 |     "#### Label Encoding"
 289 |    ]
 290 |   },
 291 |   {
 292 |    "cell_type": "code",
 293 |    "execution_count": 54,
 294 |    "metadata": {},
 295 |    "outputs": [
 296 |     {
 297 |      "name": "stdout",
 298 |      "output_type": "stream",
 299 |      "text": [
 300 |       "['Automatic' 'Manual' 'Semi-Auto']\n",
 301 |       "[' 1 Series' ' 2 Series' ' 3 Series' ' 4 Series' ' 5 Series' ' 6 Series'\n",
 302 |       " ' 7 Series' ' 8 Series' ' M2' ' M3' ' M4' ' M5' ' M6' ' X1' ' X2' ' X3'\n",
 303 |       " ' X4' ' X5' ' X6' ' X7' ' Z3' ' Z4' ' i3' ' i8']\n",
 304 |       "['Diesel' 'Electric' 'Hybrid' 'Other' 'Petrol']\n"
 305 |      ]
 306 |     },
 307 |     {
 308 |      "data": {
 309 |       "text/html": [
 310 |        "<div>\n",
 311 |        "<style scoped>\n",
 312 |        "    .dataframe tbody tr th:only-of-type {\n",
 313 |        "        vertical-align: middle;\n",
 314 |        "    }\n",
 315 |        "\n",
 316 |        "    .dataframe tbody tr th {\n",
 317 |        "        vertical-align: top;\n",
 318 |        "    }\n",
 319 |        "\n",
 320 |        "    .dataframe thead th {\n",
 321 |        "        text-align: right;\n",
 322 |        "    }\n",
 323 |        "</style>\n",
 324 |        "<table border=\"1\" class=\"dataframe\">\n",
 325 |        "  <thead>\n",
 326 |        "    <tr style=\"text-align: right;\">\n",
 327 |        "      <th></th>\n",
 328 |        "      <th>model</th>\n",
 329 |        "      <th>year</th>\n",
 330 |        "      <th>price</th>\n",
 331 |        "      <th>transmission</th>\n",
 332 |        "      <th>mileage</th>\n",
 333 |        "      <th>fuelType</th>\n",
 334 |        "      <th>tax</th>\n",
 335 |        "      <th>mpg</th>\n",
 336 |        "      <th>engineSize</th>\n",
 337 |        "      <th>transmission_</th>\n",
 338 |        "      <th>model_</th>\n",
 339 |        "      <th>fuelType_</th>\n",
 340 |        "    </tr>\n",
 341 |        "  </thead>\n",
 342 |        "  <tbody>\n",
 343 |        "    <tr>\n",
 344 |        "      <th>5852</th>\n",
 345 |        "      <td>5 Series</td>\n",
 346 |        "      <td>2017</td>\n",
 347 |        "      <td>26000</td>\n",
 348 |        "      <td>Semi-Auto</td>\n",
 349 |        "      <td>43533</td>\n",
 350 |        "      <td>Diesel</td>\n",
 351 |        "      <td>150</td>\n",
 352 |        "      <td>53.3</td>\n",
 353 |        "      <td>3.0</td>\n",
 354 |        "      <td>2</td>\n",
 355 |        "      <td>4</td>\n",
 356 |        "      <td>0</td>\n",
 357 |        "    </tr>\n",
 358 |        "    <tr>\n",
 359 |        "      <th>8204</th>\n",
 360 |        "      <td>1 Series</td>\n",
 361 |        "      <td>2017</td>\n",
 362 |        "      <td>12415</td>\n",
 363 |        "      <td>Manual</td>\n",
 364 |        "      <td>17724</td>\n",
 365 |        "      <td>Diesel</td>\n",
 366 |        "      <td>145</td>\n",
 367 |        "      <td>83.1</td>\n",
 368 |        "      <td>1.5</td>\n",
 369 |        "      <td>1</td>\n",
 370 |        "      <td>0</td>\n",
 371 |        "      <td>0</td>\n",
 372 |        "    </tr>\n",
 373 |        "    <tr>\n",
 374 |        "      <th>5318</th>\n",
 375 |        "      <td>X5</td>\n",
 376 |        "      <td>2019</td>\n",
 377 |        "      <td>52950</td>\n",
 378 |        "      <td>Semi-Auto</td>\n",
 379 |        "      <td>3309</td>\n",
 380 |        "      <td>Petrol</td>\n",
 381 |        "      <td>145</td>\n",
 382 |        "      <td>27.2</td>\n",
 383 |        "      <td>3.0</td>\n",
 384 |        "      <td>2</td>\n",
 385 |        "      <td>17</td>\n",
 386 |        "      <td>4</td>\n",
 387 |        "    </tr>\n",
 388 |        "    <tr>\n",
 389 |        "      <th>8042</th>\n",
 390 |        "      <td>4 Series</td>\n",
 391 |        "      <td>2017</td>\n",
 392 |        "      <td>19600</td>\n",
 393 |        "      <td>Automatic</td>\n",
 394 |        "      <td>19293</td>\n",
 395 |        "      <td>Diesel</td>\n",
 396 |        "      <td>145</td>\n",
 397 |        "      <td>65.7</td>\n",
 398 |        "      <td>2.0</td>\n",
 399 |        "      <td>0</td>\n",
 400 |        "      <td>3</td>\n",
 401 |        "      <td>0</td>\n",
 402 |        "    </tr>\n",
 403 |        "    <tr>\n",
 404 |        "      <th>3286</th>\n",
 405 |        "      <td>1 Series</td>\n",
 406 |        "      <td>2016</td>\n",
 407 |        "      <td>18990</td>\n",
 408 |        "      <td>Semi-Auto</td>\n",
 409 |        "      <td>11538</td>\n",
 410 |        "      <td>Petrol</td>\n",
 411 |        "      <td>200</td>\n",
 412 |        "      <td>39.8</td>\n",
 413 |        "      <td>3.0</td>\n",
 414 |        "      <td>2</td>\n",
 415 |        "      <td>0</td>\n",
 416 |        "      <td>4</td>\n",
 417 |        "    </tr>\n",
 418 |        "    <tr>\n",
 419 |        "      <th>1288</th>\n",
 420 |        "      <td>2 Series</td>\n",
 421 |        "      <td>2019</td>\n",
 422 |        "      <td>20276</td>\n",
 423 |        "      <td>Automatic</td>\n",
 424 |        "      <td>4013</td>\n",
 425 |        "      <td>Petrol</td>\n",
 426 |        "      <td>145</td>\n",
 427 |        "      <td>53.3</td>\n",
 428 |        "      <td>1.5</td>\n",
 429 |        "      <td>0</td>\n",
 430 |        "      <td>1</td>\n",
 431 |        "      <td>4</td>\n",
 432 |        "    </tr>\n",
 433 |        "    <tr>\n",
 434 |        "      <th>2329</th>\n",
 435 |        "      <td>4 Series</td>\n",
 436 |        "      <td>2018</td>\n",
 437 |        "      <td>20991</td>\n",
 438 |        "      <td>Manual</td>\n",
 439 |        "      <td>11455</td>\n",
 440 |        "      <td>Petrol</td>\n",
 441 |        "      <td>150</td>\n",
 442 |        "      <td>46.3</td>\n",
 443 |        "      <td>2.0</td>\n",
 444 |        "      <td>1</td>\n",
 445 |        "      <td>3</td>\n",
 446 |        "      <td>4</td>\n",
 447 |        "    </tr>\n",
 448 |        "    <tr>\n",
 449 |        "      <th>8830</th>\n",
 450 |        "      <td>X1</td>\n",
 451 |        "      <td>2016</td>\n",
 452 |        "      <td>17400</td>\n",
 453 |        "      <td>Manual</td>\n",
 454 |        "      <td>28999</td>\n",
 455 |        "      <td>Diesel</td>\n",
 456 |        "      <td>125</td>\n",
 457 |        "      <td>58.9</td>\n",
 458 |        "      <td>2.0</td>\n",
 459 |        "      <td>1</td>\n",
 460 |        "      <td>13</td>\n",
 461 |        "      <td>0</td>\n",
 462 |        "    </tr>\n",
 463 |        "    <tr>\n",
 464 |        "      <th>4257</th>\n",
 465 |        "      <td>2 Series</td>\n",
 466 |        "      <td>2019</td>\n",
 467 |        "      <td>21680</td>\n",
 468 |        "      <td>Semi-Auto</td>\n",
 469 |        "      <td>8994</td>\n",
 470 |        "      <td>Petrol</td>\n",
 471 |        "      <td>145</td>\n",
 472 |        "      <td>47.9</td>\n",
 473 |        "      <td>2.0</td>\n",
 474 |        "      <td>2</td>\n",
 475 |        "      <td>1</td>\n",
 476 |        "      <td>4</td>\n",
 477 |        "    </tr>\n",
 478 |        "    <tr>\n",
 479 |        "      <th>9153</th>\n",
 480 |        "      <td>Z4</td>\n",
 481 |        "      <td>2016</td>\n",
 482 |        "      <td>15991</td>\n",
 483 |        "      <td>Manual</td>\n",
 484 |        "      <td>25921</td>\n",
 485 |        "      <td>Petrol</td>\n",
 486 |        "      <td>205</td>\n",
 487 |        "      <td>41.5</td>\n",
 488 |        "      <td>2.0</td>\n",
 489 |        "      <td>1</td>\n",
 490 |        "      <td>21</td>\n",
 491 |        "      <td>4</td>\n",
 492 |        "    </tr>\n",
 493 |        "  </tbody>\n",
 494 |        "</table>\n",
 495 |        "</div>"
 496 |       ],
 497 |       "text/plain": [
 498 |        "          model  year  price transmission  mileage fuelType  tax   mpg  \\\n",
 499 |        "5852   5 Series  2017  26000    Semi-Auto    43533   Diesel  150  53.3   \n",
 500 |        "8204   1 Series  2017  12415       Manual    17724   Diesel  145  83.1   \n",
 501 |        "5318         X5  2019  52950    Semi-Auto     3309   Petrol  145  27.2   \n",
 502 |        "8042   4 Series  2017  19600    Automatic    19293   Diesel  145  65.7   \n",
 503 |        "3286   1 Series  2016  18990    Semi-Auto    11538   Petrol  200  39.8   \n",
 504 |        "1288   2 Series  2019  20276    Automatic     4013   Petrol  145  53.3   \n",
 505 |        "2329   4 Series  2018  20991       Manual    11455   Petrol  150  46.3   \n",
 506 |        "8830         X1  2016  17400       Manual    28999   Diesel  125  58.9   \n",
 507 |        "4257   2 Series  2019  21680    Semi-Auto     8994   Petrol  145  47.9   \n",
 508 |        "9153         Z4  2016  15991       Manual    25921   Petrol  205  41.5   \n",
 509 |        "\n",
 510 |        "      engineSize  transmission_  model_  fuelType_  \n",
 511 |        "5852         3.0              2       4          0  \n",
 512 |        "8204         1.5              1       0          0  \n",
 513 |        "5318         3.0              2      17          4  \n",
 514 |        "8042         2.0              0       3          0  \n",
 515 |        "3286         3.0              2       0          4  \n",
 516 |        "1288         1.5              0       1          4  \n",
 517 |        "2329         2.0              1       3          4  \n",
 518 |        "8830         2.0              1      13          0  \n",
 519 |        "4257         2.0              2       1          4  \n",
 520 |        "9153         2.0              1      21          4  "
 521 |       ]
 522 |      },
 523 |      "metadata": {},
 524 |      "output_type": "display_data"
 525 |     }
 526 |    ],
 527 |    "source": [
 528 |     "LE = LabelEncoder()\n",
 529 |     "df[\"transmission_\"] = LE.fit_transform(df[\"transmission\"])\n",
 530 |     "print(LE.classes_)\n",
 531 |     "# print(df[\"transmission_\"].unique())\n",
 532 |     "\n",
 533 |     "df[\"model_\"] = LE.fit_transform(df[\"model\"])\n",
 534 |     "print(LE.classes_)\n",
 535 |     "# print(df[\"model_\"].unique())\n",
 536 |     "\n",
 537 |     "df[\"fuelType_\"] = LE.fit_transform(df[\"fuelType\"])\n",
 538 |     "print(LE.classes_)\n",
 539 |     "# print(df[\"fuelType_\"].unique())\n",
 540 |     "\n",
 541 |     "display(df.sample(10))"
 542 |    ]
 543 |   },
 544 |   {
 545 |    "cell_type": "code",
 546 |    "execution_count": 55,
 547 |    "metadata": {},
 548 |    "outputs": [
 549 |     {
 550 |      "name": "stdout",
 551 |      "output_type": "stream",
 552 |      "text": [
 553 |       "<class 'pandas.core.frame.DataFrame'>\n",
 554 |       "RangeIndex: 10781 entries, 0 to 10780\n",
 555 |       "Data columns (total 12 columns):\n",
 556 |       " #   Column         Non-Null Count  Dtype  \n",
 557 |       "---  ------         --------------  -----  \n",
 558 |       " 0   model          10781 non-null  object \n",
 559 |       " 1   year           10781 non-null  int64  \n",
 560 |       " 2   price          10781 non-null  int64  \n",
 561 |       " 3   transmission   10781 non-null  object \n",
 562 |       " 4   mileage        10781 non-null  int64  \n",
 563 |       " 5   fuelType       10781 non-null  object \n",
 564 |       " 6   tax            10781 non-null  int64  \n",
 565 |       " 7   mpg            10781 non-null  float64\n",
 566 |       " 8   engineSize     10781 non-null  float64\n",
 567 |       " 9   transmission_  10781 non-null  int64  \n",
 568 |       " 10  model_         10781 non-null  int64  \n",
 569 |       " 11  fuelType_      10781 non-null  int64  \n",
 570 |       "dtypes: float64(2), int64(7), object(3)\n",
 571 |       "memory usage: 1010.8+ KB\n"
 572 |      ]
 573 |     }
 574 |    ],
 575 |    "source": [
 576 |     "df.info()"
 577 |    ]
 578 |   },
 579 |   {
 580 |    "cell_type": "code",
 581 |    "execution_count": 56,
 582 |    "metadata": {},
 583 |    "outputs": [
 584 |     {
 585 |      "data": {
 586 |       "text/plain": [
 587 |        "Semi-Auto    4666\n",
 588 |        "Automatic    3588\n",
 589 |        "Manual       2527\n",
 590 |        "Name: transmission, dtype: int64"
 591 |       ]
 592 |      },
 593 |      "execution_count": 56,
 594 |      "metadata": {},
 595 |      "output_type": "execute_result"
 596 |     }
 597 |    ],
 598 |    "source": [
 599 |     "df[\"transmission\"].value_counts()"
 600 |    ]
 601 |   },
 602 |   {
 603 |    "cell_type": "code",
 604 |    "execution_count": 57,
 605 |    "metadata": {},
 606 |    "outputs": [
 607 |     {
 608 |      "data": {
 609 |       "text/plain": [
 610 |        "Diesel      7027\n",
 611 |        "Petrol      3417\n",
 612 |        "Hybrid       298\n",
 613 |        "Other         36\n",
 614 |        "Electric       3\n",
 615 |        "Name: fuelType, dtype: int64"
 616 |       ]
 617 |      },
 618 |      "execution_count": 57,
 619 |      "metadata": {},
 620 |      "output_type": "execute_result"
 621 |     }
 622 |    ],
 623 |    "source": [
 624 |     "df[\"fuelType\"].value_counts()"
 625 |    ]
 626 |   },
 627 |   {
 628 |    "cell_type": "markdown",
 629 |    "metadata": {},
 630 |    "source": [
 631 |     "#### One-hot Encoding"
 632 |    ]
 633 |   },
 634 |   {
 635 |    "cell_type": "code",
 636 |    "execution_count": 58,
 637 |    "metadata": {},
 638 |    "outputs": [
 639 |     {
 640 |      "data": {
 641 |       "text/html": [
 642 |        "<div>\n",
 643 |        "<style scoped>\n",
 644 |        "    .dataframe tbody tr th:only-of-type {\n",
 645 |        "        vertical-align: middle;\n",
 646 |        "    }\n",
 647 |        "\n",
 648 |        "    .dataframe tbody tr th {\n",
 649 |        "        vertical-align: top;\n",
 650 |        "    }\n",
 651 |        "\n",
 652 |        "    .dataframe thead th {\n",
 653 |        "        text-align: right;\n",
 654 |        "    }\n",
 655 |        "</style>\n",
 656 |        "<table border=\"1\" class=\"dataframe\">\n",
 657 |        "  <thead>\n",
 658 |        "    <tr style=\"text-align: right;\">\n",
 659 |        "      <th></th>\n",
 660 |        "      <th>Automatic</th>\n",
 661 |        "      <th>Semi-Auto</th>\n",
 662 |        "    </tr>\n",
 663 |        "  </thead>\n",
 664 |        "  <tbody>\n",
 665 |        "    <tr>\n",
 666 |        "      <th>5989</th>\n",
 667 |        "      <td>0</td>\n",
 668 |        "      <td>1</td>\n",
 669 |        "    </tr>\n",
 670 |        "    <tr>\n",
 671 |        "      <th>2891</th>\n",
 672 |        "      <td>0</td>\n",
 673 |        "      <td>1</td>\n",
 674 |        "    </tr>\n",
 675 |        "    <tr>\n",
 676 |        "      <th>9332</th>\n",
 677 |        "      <td>1</td>\n",
 678 |        "      <td>0</td>\n",
 679 |        "    </tr>\n",
 680 |        "    <tr>\n",
 681 |        "      <th>7102</th>\n",
 682 |        "      <td>1</td>\n",
 683 |        "      <td>0</td>\n",
 684 |        "    </tr>\n",
 685 |        "    <tr>\n",
 686 |        "      <th>3650</th>\n",
 687 |        "      <td>0</td>\n",
 688 |        "      <td>0</td>\n",
 689 |        "    </tr>\n",
 690 |        "    <tr>\n",
 691 |        "      <th>4266</th>\n",
 692 |        "      <td>0</td>\n",
 693 |        "      <td>0</td>\n",
 694 |        "    </tr>\n",
 695 |        "    <tr>\n",
 696 |        "      <th>1409</th>\n",
 697 |        "      <td>1</td>\n",
 698 |        "      <td>0</td>\n",
 699 |        "    </tr>\n",
 700 |        "    <tr>\n",
 701 |        "      <th>505</th>\n",
 702 |        "      <td>0</td>\n",
 703 |        "      <td>1</td>\n",
 704 |        "    </tr>\n",
 705 |        "    <tr>\n",
 706 |        "      <th>9112</th>\n",
 707 |        "      <td>1</td>\n",
 708 |        "      <td>0</td>\n",
 709 |        "    </tr>\n",
 710 |        "    <tr>\n",
 711 |        "      <th>6345</th>\n",
 712 |        "      <td>0</td>\n",
 713 |        "      <td>1</td>\n",
 714 |        "    </tr>\n",
 715 |        "  </tbody>\n",
 716 |        "</table>\n",
 717 |        "</div>"
 718 |       ],
 719 |       "text/plain": [
 720 |        "      Automatic  Semi-Auto\n",
 721 |        "5989          0          1\n",
 722 |        "2891          0          1\n",
 723 |        "9332          1          0\n",
 724 |        "7102          1          0\n",
 725 |        "3650          0          0\n",
 726 |        "4266          0          0\n",
 727 |        "1409          1          0\n",
 728 |        "505           0          1\n",
 729 |        "9112          1          0\n",
 730 |        "6345          0          1"
 731 |       ]
 732 |      },
 733 |      "metadata": {},
 734 |      "output_type": "display_data"
 735 |     }
 736 |    ],
 737 |    "source": [
 738 |     "df_transmission = pd.get_dummies(df[\"transmission\"])\n",
 739 |     "df_transmission = df_transmission.drop(\"Manual\", axis=1)\n",
 740 |     "display(df_transmission.sample(10))"
 741 |    ]
 742 |   },
 743 |   {
 744 |    "cell_type": "code",
 745 |    "execution_count": 59,
 746 |    "metadata": {},
 747 |    "outputs": [
 748 |     {
 749 |      "data": {
 750 |       "text/html": [
 751 |        "<div>\n",
 752 |        "<style scoped>\n",
 753 |        "    .dataframe tbody tr th:only-of-type {\n",
 754 |        "        vertical-align: middle;\n",
 755 |        "    }\n",
 756 |        "\n",
 757 |        "    .dataframe tbody tr th {\n",
 758 |        "        vertical-align: top;\n",
 759 |        "    }\n",
 760 |        "\n",
 761 |        "    .dataframe thead th {\n",
 762 |        "        text-align: right;\n",
 763 |        "    }\n",
 764 |        "</style>\n",
 765 |        "<table border=\"1\" class=\"dataframe\">\n",
 766 |        "  <thead>\n",
 767 |        "    <tr style=\"text-align: right;\">\n",
 768 |        "      <th></th>\n",
 769 |        "      <th>Diesel</th>\n",
 770 |        "      <th>Electric</th>\n",
 771 |        "      <th>Hybrid</th>\n",
 772 |        "      <th>Other</th>\n",
 773 |        "      <th>Petrol</th>\n",
 774 |        "    </tr>\n",
 775 |        "  </thead>\n",
 776 |        "  <tbody>\n",
 777 |        "    <tr>\n",
 778 |        "      <th>917</th>\n",
 779 |        "      <td>1</td>\n",
 780 |        "      <td>0</td>\n",
 781 |        "      <td>0</td>\n",
 782 |        "      <td>0</td>\n",
 783 |        "      <td>0</td>\n",
 784 |        "    </tr>\n",
 785 |        "    <tr>\n",
 786 |        "      <th>4115</th>\n",
 787 |        "      <td>0</td>\n",
 788 |        "      <td>0</td>\n",
 789 |        "      <td>0</td>\n",
 790 |        "      <td>0</td>\n",
 791 |        "      <td>1</td>\n",
 792 |        "    </tr>\n",
 793 |        "    <tr>\n",
 794 |        "      <th>3506</th>\n",
 795 |        "      <td>1</td>\n",
 796 |        "      <td>0</td>\n",
 797 |        "      <td>0</td>\n",
 798 |        "      <td>0</td>\n",
 799 |        "      <td>0</td>\n",
 800 |        "    </tr>\n",
 801 |        "    <tr>\n",
 802 |        "      <th>4660</th>\n",
 803 |        "      <td>1</td>\n",
 804 |        "      <td>0</td>\n",
 805 |        "      <td>0</td>\n",
 806 |        "      <td>0</td>\n",
 807 |        "      <td>0</td>\n",
 808 |        "    </tr>\n",
 809 |        "    <tr>\n",
 810 |        "      <th>6027</th>\n",
 811 |        "      <td>0</td>\n",
 812 |        "      <td>0</td>\n",
 813 |        "      <td>0</td>\n",
 814 |        "      <td>0</td>\n",
 815 |        "      <td>1</td>\n",
 816 |        "    </tr>\n",
 817 |        "    <tr>\n",
 818 |        "      <th>9218</th>\n",
 819 |        "      <td>0</td>\n",
 820 |        "      <td>0</td>\n",
 821 |        "      <td>0</td>\n",
 822 |        "      <td>0</td>\n",
 823 |        "      <td>1</td>\n",
 824 |        "    </tr>\n",
 825 |        "    <tr>\n",
 826 |        "      <th>7249</th>\n",
 827 |        "      <td>0</td>\n",
 828 |        "      <td>0</td>\n",
 829 |        "      <td>0</td>\n",
 830 |        "      <td>0</td>\n",
 831 |        "      <td>1</td>\n",
 832 |        "    </tr>\n",
 833 |        "    <tr>\n",
 834 |        "      <th>7939</th>\n",
 835 |        "      <td>1</td>\n",
 836 |        "      <td>0</td>\n",
 837 |        "      <td>0</td>\n",
 838 |        "      <td>0</td>\n",
 839 |        "      <td>0</td>\n",
 840 |        "    </tr>\n",
 841 |        "    <tr>\n",
 842 |        "      <th>4755</th>\n",
 843 |        "      <td>1</td>\n",
 844 |        "      <td>0</td>\n",
 845 |        "      <td>0</td>\n",
 846 |        "      <td>0</td>\n",
 847 |        "      <td>0</td>\n",
 848 |        "    </tr>\n",
 849 |        "    <tr>\n",
 850 |        "      <th>2139</th>\n",
 851 |        "      <td>0</td>\n",
 852 |        "      <td>0</td>\n",
 853 |        "      <td>0</td>\n",
 854 |        "      <td>0</td>\n",
 855 |        "      <td>1</td>\n",
 856 |        "    </tr>\n",
 857 |        "  </tbody>\n",
 858 |        "</table>\n",
 859 |        "</div>"
 860 |       ],
 861 |       "text/plain": [
 862 |        "      Diesel  Electric  Hybrid  Other  Petrol\n",
 863 |        "917        1         0       0      0       0\n",
 864 |        "4115       0         0       0      0       1\n",
 865 |        "3506       1         0       0      0       0\n",
 866 |        "4660       1         0       0      0       0\n",
 867 |        "6027       0         0       0      0       1\n",
 868 |        "9218       0         0       0      0       1\n",
 869 |        "7249       0         0       0      0       1\n",
 870 |        "7939       1         0       0      0       0\n",
 871 |        "4755       1         0       0      0       0\n",
 872 |        "2139       0         0       0      0       1"
 873 |       ]
 874 |      },
 875 |      "metadata": {},
 876 |      "output_type": "display_data"
 877 |     }
 878 |    ],
 879 |    "source": [
 880 |     "df_fuelType = pd.get_dummies(df[\"fuelType\"])\n",
 881 |     "display(df_fuelType.sample(10))"
 882 |    ]
 883 |   },
 884 |   {
 885 |    "cell_type": "code",
 886 |    "execution_count": 60,
 887 |    "metadata": {},
 888 |    "outputs": [
 889 |     {
 890 |      "data": {
 891 |       "text/html": [
 892 |        "<div>\n",
 893 |        "<style scoped>\n",
 894 |        "    .dataframe tbody tr th:only-of-type {\n",
 895 |        "        vertical-align: middle;\n",
 896 |        "    }\n",
 897 |        "\n",
 898 |        "    .dataframe tbody tr th {\n",
 899 |        "        vertical-align: top;\n",
 900 |        "    }\n",
 901 |        "\n",
 902 |        "    .dataframe thead th {\n",
 903 |        "        text-align: right;\n",
 904 |        "    }\n",
 905 |        "</style>\n",
 906 |        "<table border=\"1\" class=\"dataframe\">\n",
 907 |        "  <thead>\n",
 908 |        "    <tr style=\"text-align: right;\">\n",
 909 |        "      <th></th>\n",
 910 |        "      <th>1 Series</th>\n",
 911 |        "      <th>2 Series</th>\n",
 912 |        "      <th>3 Series</th>\n",
 913 |        "      <th>4 Series</th>\n",
 914 |        "      <th>5 Series</th>\n",
 915 |        "      <th>6 Series</th>\n",
 916 |        "      <th>7 Series</th>\n",
 917 |        "      <th>8 Series</th>\n",
 918 |        "      <th>M2</th>\n",
 919 |        "      <th>M3</th>\n",
 920 |        "      <th>...</th>\n",
 921 |        "      <th>X2</th>\n",
 922 |        "      <th>X3</th>\n",
 923 |        "      <th>X4</th>\n",
 924 |        "      <th>X5</th>\n",
 925 |        "      <th>X6</th>\n",
 926 |        "      <th>X7</th>\n",
 927 |        "      <th>Z3</th>\n",
 928 |        "      <th>Z4</th>\n",
 929 |        "      <th>i3</th>\n",
 930 |        "      <th>i8</th>\n",
 931 |        "    </tr>\n",
 932 |        "  </thead>\n",
 933 |        "  <tbody>\n",
 934 |        "    <tr>\n",
 935 |        "      <th>10633</th>\n",
 936 |        "      <td>0</td>\n",
 937 |        "      <td>0</td>\n",
 938 |        "      <td>0</td>\n",
 939 |        "      <td>0</td>\n",
 940 |        "      <td>1</td>\n",
 941 |        "      <td>0</td>\n",
 942 |        "      <td>0</td>\n",
 943 |        "      <td>0</td>\n",
 944 |        "      <td>0</td>\n",
 945 |        "      <td>0</td>\n",
 946 |        "      <td>...</td>\n",
 947 |        "      <td>0</td>\n",
 948 |        "      <td>0</td>\n",
 949 |        "      <td>0</td>\n",
 950 |        "      <td>0</td>\n",
 951 |        "      <td>0</td>\n",
 952 |        "      <td>0</td>\n",
 953 |        "      <td>0</td>\n",
 954 |        "      <td>0</td>\n",
 955 |        "      <td>0</td>\n",
 956 |        "      <td>0</td>\n",
 957 |        "    </tr>\n",
 958 |        "    <tr>\n",
 959 |        "      <th>5592</th>\n",
 960 |        "      <td>0</td>\n",
 961 |        "      <td>1</td>\n",
 962 |        "      <td>0</td>\n",
 963 |        "      <td>0</td>\n",
 964 |        "      <td>0</td>\n",
 965 |        "      <td>0</td>\n",
 966 |        "      <td>0</td>\n",
 967 |        "      <td>0</td>\n",
 968 |        "      <td>0</td>\n",
 969 |        "      <td>0</td>\n",
 970 |        "      <td>...</td>\n",
 971 |        "      <td>0</td>\n",
 972 |        "      <td>0</td>\n",
 973 |        "      <td>0</td>\n",
 974 |        "      <td>0</td>\n",
 975 |        "      <td>0</td>\n",
 976 |        "      <td>0</td>\n",
 977 |        "      <td>0</td>\n",
 978 |        "      <td>0</td>\n",
 979 |        "      <td>0</td>\n",
 980 |        "      <td>0</td>\n",
 981 |        "    </tr>\n",
 982 |        "    <tr>\n",
 983 |        "      <th>10773</th>\n",
 984 |        "      <td>0</td>\n",
 985 |        "      <td>0</td>\n",
 986 |        "      <td>0</td>\n",
 987 |        "      <td>0</td>\n",
 988 |        "      <td>0</td>\n",
 989 |        "      <td>0</td>\n",
 990 |        "      <td>0</td>\n",
 991 |        "      <td>0</td>\n",
 992 |        "      <td>0</td>\n",
 993 |        "      <td>0</td>\n",
 994 |        "      <td>...</td>\n",
 995 |        "      <td>0</td>\n",
 996 |        "      <td>0</td>\n",
 997 |        "      <td>0</td>\n",
 998 |        "      <td>0</td>\n",
 999 |        "      <td>0</td>\n",
1000 |        "      <td>0</td>\n",
1001 |        "      <td>0</td>\n",
1002 |        "      <td>0</td>\n",
1003 |        "      <td>0</td>\n",
1004 |        "      <td>0</td>\n",
1005 |        "    </tr>\n",
1006 |        "    <tr>\n",
1007 |        "      <th>5449</th>\n",
1008 |        "      <td>0</td>\n",
1009 |        "      <td>0</td>\n",
1010 |        "      <td>1</td>\n",
1011 |        "      <td>0</td>\n",
1012 |        "      <td>0</td>\n",
1013 |        "      <td>0</td>\n",
1014 |        "      <td>0</td>\n",
1015 |        "      <td>0</td>\n",
1016 |        "      <td>0</td>\n",
1017 |        "      <td>0</td>\n",
1018 |        "      <td>...</td>\n",
1019 |        "      <td>0</td>\n",
1020 |        "      <td>0</td>\n",
1021 |        "      <td>0</td>\n",
1022 |        "      <td>0</td>\n",
1023 |        "      <td>0</td>\n",
1024 |        "      <td>0</td>\n",
1025 |        "      <td>0</td>\n",
1026 |        "      <td>0</td>\n",
1027 |        "      <td>0</td>\n",
1028 |        "      <td>0</td>\n",
1029 |        "    </tr>\n",
1030 |        "    <tr>\n",
1031 |        "      <th>8481</th>\n",
1032 |        "      <td>0</td>\n",
1033 |        "      <td>1</td>\n",
1034 |        "      <td>0</td>\n",
1035 |        "      <td>0</td>\n",
1036 |        "      <td>0</td>\n",
1037 |        "      <td>0</td>\n",
1038 |        "      <td>0</td>\n",
1039 |        "      <td>0</td>\n",
1040 |        "      <td>0</td>\n",
1041 |        "      <td>0</td>\n",
1042 |        "      <td>...</td>\n",
1043 |        "      <td>0</td>\n",
1044 |        "      <td>0</td>\n",
1045 |        "      <td>0</td>\n",
1046 |        "      <td>0</td>\n",
1047 |        "      <td>0</td>\n",
1048 |        "      <td>0</td>\n",
1049 |        "      <td>0</td>\n",
1050 |        "      <td>0</td>\n",
1051 |        "      <td>0</td>\n",
1052 |        "      <td>0</td>\n",
1053 |        "    </tr>\n",
1054 |        "    <tr>\n",
1055 |        "      <th>8746</th>\n",
1056 |        "      <td>0</td>\n",
1057 |        "      <td>0</td>\n",
1058 |        "      <td>1</td>\n",
1059 |        "      <td>0</td>\n",
1060 |        "      <td>0</td>\n",
1061 |        "      <td>0</td>\n",
1062 |        "      <td>0</td>\n",
1063 |        "      <td>0</td>\n",
1064 |        "      <td>0</td>\n",
1065 |        "      <td>0</td>\n",
1066 |        "      <td>...</td>\n",
1067 |        "      <td>0</td>\n",
1068 |        "      <td>0</td>\n",
1069 |        "      <td>0</td>\n",
1070 |        "      <td>0</td>\n",
1071 |        "      <td>0</td>\n",
1072 |        "      <td>0</td>\n",
1073 |        "      <td>0</td>\n",
1074 |        "      <td>0</td>\n",
1075 |        "      <td>0</td>\n",
1076 |        "      <td>0</td>\n",
1077 |        "    </tr>\n",
1078 |        "    <tr>\n",
1079 |        "      <th>2061</th>\n",
1080 |        "      <td>0</td>\n",
1081 |        "      <td>0</td>\n",
1082 |        "      <td>0</td>\n",
1083 |        "      <td>1</td>\n",
1084 |        "      <td>0</td>\n",
1085 |        "      <td>0</td>\n",
1086 |        "      <td>0</td>\n",
1087 |        "      <td>0</td>\n",
1088 |        "      <td>0</td>\n",
1089 |        "      <td>0</td>\n",
1090 |        "      <td>...</td>\n",
1091 |        "      <td>0</td>\n",
1092 |        "      <td>0</td>\n",
1093 |        "      <td>0</td>\n",
1094 |        "      <td>0</td>\n",
1095 |        "      <td>0</td>\n",
1096 |        "      <td>0</td>\n",
1097 |        "      <td>0</td>\n",
1098 |        "      <td>0</td>\n",
1099 |        "      <td>0</td>\n",
1100 |        "      <td>0</td>\n",
1101 |        "    </tr>\n",
1102 |        "    <tr>\n",
1103 |        "      <th>5764</th>\n",
1104 |        "      <td>1</td>\n",
1105 |        "      <td>0</td>\n",
1106 |        "      <td>0</td>\n",
1107 |        "      <td>0</td>\n",
1108 |        "      <td>0</td>\n",
1109 |        "      <td>0</td>\n",
1110 |        "      <td>0</td>\n",
1111 |        "      <td>0</td>\n",
1112 |        "      <td>0</td>\n",
1113 |        "      <td>0</td>\n",
1114 |        "      <td>...</td>\n",
1115 |        "      <td>0</td>\n",
1116 |        "      <td>0</td>\n",
1117 |        "      <td>0</td>\n",
1118 |        "      <td>0</td>\n",
1119 |        "      <td>0</td>\n",
1120 |        "      <td>0</td>\n",
1121 |        "      <td>0</td>\n",
1122 |        "      <td>0</td>\n",
1123 |        "      <td>0</td>\n",
1124 |        "      <td>0</td>\n",
1125 |        "    </tr>\n",
1126 |        "    <tr>\n",
1127 |        "      <th>9317</th>\n",
1128 |        "      <td>0</td>\n",
1129 |        "      <td>0</td>\n",
1130 |        "      <td>1</td>\n",
1131 |        "      <td>0</td>\n",
1132 |        "      <td>0</td>\n",
1133 |        "      <td>0</td>\n",
1134 |        "      <td>0</td>\n",
1135 |        "      <td>0</td>\n",
1136 |        "      <td>0</td>\n",
1137 |        "      <td>0</td>\n",
1138 |        "      <td>...</td>\n",
1139 |        "      <td>0</td>\n",
1140 |        "      <td>0</td>\n",
1141 |        "      <td>0</td>\n",
1142 |        "      <td>0</td>\n",
1143 |        "      <td>0</td>\n",
1144 |        "      <td>0</td>\n",
1145 |        "      <td>0</td>\n",
1146 |        "      <td>0</td>\n",
1147 |        "      <td>0</td>\n",
1148 |        "      <td>0</td>\n",
1149 |        "    </tr>\n",
1150 |        "    <tr>\n",
1151 |        "      <th>1761</th>\n",
1152 |        "      <td>0</td>\n",
1153 |        "      <td>0</td>\n",
1154 |        "      <td>1</td>\n",
1155 |        "      <td>0</td>\n",
1156 |        "      <td>0</td>\n",
1157 |        "      <td>0</td>\n",
1158 |        "      <td>0</td>\n",
1159 |        "      <td>0</td>\n",
1160 |        "      <td>0</td>\n",
1161 |        "      <td>0</td>\n",
1162 |        "      <td>...</td>\n",
1163 |        "      <td>0</td>\n",
1164 |        "      <td>0</td>\n",
1165 |        "      <td>0</td>\n",
1166 |        "      <td>0</td>\n",
1167 |        "      <td>0</td>\n",
1168 |        "      <td>0</td>\n",
1169 |        "      <td>0</td>\n",
1170 |        "      <td>0</td>\n",
1171 |        "      <td>0</td>\n",
1172 |        "      <td>0</td>\n",
1173 |        "    </tr>\n",
1174 |        "  </tbody>\n",
1175 |        "</table>\n",
1176 |        "<p>10 rows × 24 columns</p>\n",
1177 |        "</div>"
1178 |       ],
1179 |       "text/plain": [
1180 |        "        1 Series   2 Series   3 Series   4 Series   5 Series   6 Series  \\\n",
1181 |        "10633          0          0          0          0          1          0   \n",
1182 |        "5592           0          1          0          0          0          0   \n",
1183 |        "10773          0          0          0          0          0          0   \n",
1184 |        "5449           0          0          1          0          0          0   \n",
1185 |        "8481           0          1          0          0          0          0   \n",
1186 |        "8746           0          0          1          0          0          0   \n",
1187 |        "2061           0          0          0          1          0          0   \n",
1188 |        "5764           1          0          0          0          0          0   \n",
1189 |        "9317           0          0          1          0          0          0   \n",
1190 |        "1761           0          0          1          0          0          0   \n",
1191 |        "\n",
1192 |        "        7 Series   8 Series   M2   M3  ...   X2   X3   X4   X5   X6   X7   Z3  \\\n",
1193 |        "10633          0          0    0    0  ...    0    0    0    0    0    0    0   \n",
1194 |        "5592           0          0    0    0  ...    0    0    0    0    0    0    0   \n",
1195 |        "10773          0          0    0    0  ...    0    0    0    0    0    0    0   \n",
1196 |        "5449           0          0    0    0  ...    0    0    0    0    0    0    0   \n",
1197 |        "8481           0          0    0    0  ...    0    0    0    0    0    0    0   \n",
1198 |        "8746           0          0    0    0  ...    0    0    0    0    0    0    0   \n",
1199 |        "2061           0          0    0    0  ...    0    0    0    0    0    0    0   \n",
1200 |        "5764           0          0    0    0  ...    0    0    0    0    0    0    0   \n",
1201 |        "9317           0          0    0    0  ...    0    0    0    0    0    0    0   \n",
1202 |        "1761           0          0    0    0  ...    0    0    0    0    0    0    0   \n",
1203 |        "\n",
1204 |        "        Z4   i3   i8  \n",
1205 |        "10633    0    0    0  \n",
1206 |        "5592     0    0    0  \n",
1207 |        "10773    0    0    0  \n",
1208 |        "5449     0    0    0  \n",
1209 |        "8481     0    0    0  \n",
1210 |        "8746     0    0    0  \n",
1211 |        "2061     0    0    0  \n",
1212 |        "5764     0    0    0  \n",
1213 |        "9317     0    0    0  \n",
1214 |        "1761     0    0    0  \n",
1215 |        "\n",
1216 |        "[10 rows x 24 columns]"
1217 |       ]
1218 |      },
1219 |      "metadata": {},
1220 |      "output_type": "display_data"
1221 |     }
1222 |    ],
1223 |    "source": [
1224 |     "df_model = pd.get_dummies(df[\"model\"])\n",
1225 |     "display(df_model.sample(10))"
1226 |    ]
1227 |   },
1228 |   {
1229 |    "cell_type": "code",
1230 |    "execution_count": 61,
1231 |    "metadata": {},
1232 |    "outputs": [
1233 |     {
1234 |      "name": "stdout",
1235 |      "output_type": "stream",
1236 |      "text": [
1237 |       "(10781, 12)\n"
1238 |      ]
1239 |     },
1240 |     {
1241 |      "data": {
1242 |       "text/html": [
1243 |        "<div>\n",
1244 |        "<style scoped>\n",
1245 |        "    .dataframe tbody tr th:only-of-type {\n",
1246 |        "        vertical-align: middle;\n",
1247 |        "    }\n",
1248 |        "\n",
1249 |        "    .dataframe tbody tr th {\n",
1250 |        "        vertical-align: top;\n",
1251 |        "    }\n",
1252 |        "\n",
1253 |        "    .dataframe thead th {\n",
1254 |        "        text-align: right;\n",
1255 |        "    }\n",
1256 |        "</style>\n",
1257 |        "<table border=\"1\" class=\"dataframe\">\n",
1258 |        "  <thead>\n",
1259 |        "    <tr style=\"text-align: right;\">\n",
1260 |        "      <th></th>\n",
1261 |        "      <th>year</th>\n",
1262 |        "      <th>price</th>\n",
1263 |        "      <th>mileage</th>\n",
1264 |        "      <th>tax</th>\n",
1265 |        "      <th>mpg</th>\n",
1266 |        "      <th>engineSize</th>\n",
1267 |        "      <th>transmission_</th>\n",
1268 |        "      <th>model_</th>\n",
1269 |        "      <th>fuelType_</th>\n",
1270 |        "      <th>model_ 1 Series</th>\n",
1271 |        "      <th>...</th>\n",
1272 |        "      <th>model_ i3</th>\n",
1273 |        "      <th>model_ i8</th>\n",
1274 |        "      <th>transmission_Automatic</th>\n",
1275 |        "      <th>transmission_Manual</th>\n",
1276 |        "      <th>transmission_Semi-Auto</th>\n",
1277 |        "      <th>fuelType_Diesel</th>\n",
1278 |        "      <th>fuelType_Electric</th>\n",
1279 |        "      <th>fuelType_Hybrid</th>\n",
1280 |        "      <th>fuelType_Other</th>\n",
1281 |        "      <th>fuelType_Petrol</th>\n",
1282 |        "    </tr>\n",
1283 |        "  </thead>\n",
1284 |        "  <tbody>\n",
1285 |        "    <tr>\n",
1286 |        "      <th>646</th>\n",
1287 |        "      <td>2019</td>\n",
1288 |        "      <td>29998</td>\n",
1289 |        "      <td>10171</td>\n",
1290 |        "      <td>145</td>\n",
1291 |        "      <td>41.5</td>\n",
1292 |        "      <td>2.0</td>\n",
1293 |        "      <td>2</td>\n",
1294 |        "      <td>2</td>\n",
1295 |        "      <td>4</td>\n",
1296 |        "      <td>0</td>\n",
1297 |        "      <td>...</td>\n",
1298 |        "      <td>0</td>\n",
1299 |        "      <td>0</td>\n",
1300 |        "      <td>0</td>\n",
1301 |        "      <td>0</td>\n",
1302 |        "      <td>1</td>\n",
1303 |        "      <td>0</td>\n",
1304 |        "      <td>0</td>\n",
1305 |        "      <td>0</td>\n",
1306 |        "      <td>0</td>\n",
1307 |        "      <td>1</td>\n",
1308 |        "    </tr>\n",
1309 |        "    <tr>\n",
1310 |        "      <th>9818</th>\n",
1311 |        "      <td>2016</td>\n",
1312 |        "      <td>17999</td>\n",
1313 |        "      <td>27193</td>\n",
1314 |        "      <td>160</td>\n",
1315 |        "      <td>50.4</td>\n",
1316 |        "      <td>3.0</td>\n",
1317 |        "      <td>0</td>\n",
1318 |        "      <td>4</td>\n",
1319 |        "      <td>0</td>\n",
1320 |        "      <td>0</td>\n",
1321 |        "      <td>...</td>\n",
1322 |        "      <td>0</td>\n",
1323 |        "      <td>0</td>\n",
1324 |        "      <td>1</td>\n",
1325 |        "      <td>0</td>\n",
1326 |        "      <td>0</td>\n",
1327 |        "      <td>1</td>\n",
1328 |        "      <td>0</td>\n",
1329 |        "      <td>0</td>\n",
1330 |        "      <td>0</td>\n",
1331 |        "      <td>0</td>\n",
1332 |        "    </tr>\n",
1333 |        "    <tr>\n",
1334 |        "      <th>8723</th>\n",
1335 |        "      <td>2017</td>\n",
1336 |        "      <td>13000</td>\n",
1337 |        "      <td>24104</td>\n",
1338 |        "      <td>125</td>\n",
1339 |        "      <td>53.3</td>\n",
1340 |        "      <td>1.5</td>\n",
1341 |        "      <td>1</td>\n",
1342 |        "      <td>0</td>\n",
1343 |        "      <td>4</td>\n",
1344 |        "      <td>1</td>\n",
1345 |        "      <td>...</td>\n",
1346 |        "      <td>0</td>\n",
1347 |        "      <td>0</td>\n",
1348 |        "      <td>0</td>\n",
1349 |        "      <td>1</td>\n",
1350 |        "      <td>0</td>\n",
1351 |        "      <td>0</td>\n",
1352 |        "      <td>0</td>\n",
1353 |        "      <td>0</td>\n",
1354 |        "      <td>0</td>\n",
1355 |        "      <td>1</td>\n",
1356 |        "    </tr>\n",
1357 |        "    <tr>\n",
1358 |        "      <th>3323</th>\n",
1359 |        "      <td>2019</td>\n",
1360 |        "      <td>50990</td>\n",
1361 |        "      <td>4729</td>\n",
1362 |        "      <td>145</td>\n",
1363 |        "      <td>37.7</td>\n",
1364 |        "      <td>3.0</td>\n",
1365 |        "      <td>2</td>\n",
1366 |        "      <td>17</td>\n",
1367 |        "      <td>0</td>\n",
1368 |        "      <td>0</td>\n",
1369 |        "      <td>...</td>\n",
1370 |        "      <td>0</td>\n",
1371 |        "      <td>0</td>\n",
1372 |        "      <td>0</td>\n",
1373 |        "      <td>0</td>\n",
1374 |        "      <td>1</td>\n",
1375 |        "      <td>1</td>\n",
1376 |        "      <td>0</td>\n",
1377 |        "      <td>0</td>\n",
1378 |        "      <td>0</td>\n",
1379 |        "      <td>0</td>\n",
1380 |        "    </tr>\n",
1381 |        "    <tr>\n",
1382 |        "      <th>2641</th>\n",
1383 |        "      <td>2019</td>\n",
1384 |        "      <td>30326</td>\n",
1385 |        "      <td>3392</td>\n",
1386 |        "      <td>145</td>\n",
1387 |        "      <td>41.5</td>\n",
1388 |        "      <td>2.0</td>\n",
1389 |        "      <td>2</td>\n",
1390 |        "      <td>2</td>\n",
1391 |        "      <td>4</td>\n",
1392 |        "      <td>0</td>\n",
1393 |        "      <td>...</td>\n",
1394 |        "      <td>0</td>\n",
1395 |        "      <td>0</td>\n",
1396 |        "      <td>0</td>\n",
1397 |        "      <td>0</td>\n",
1398 |        "      <td>1</td>\n",
1399 |        "      <td>0</td>\n",
1400 |        "      <td>0</td>\n",
1401 |        "      <td>0</td>\n",
1402 |        "      <td>0</td>\n",
1403 |        "      <td>1</td>\n",
1404 |        "    </tr>\n",
1405 |        "    <tr>\n",
1406 |        "      <th>2205</th>\n",
1407 |        "      <td>2015</td>\n",
1408 |        "      <td>10999</td>\n",
1409 |        "      <td>90018</td>\n",
1410 |        "      <td>20</td>\n",
1411 |        "      <td>70.6</td>\n",
1412 |        "      <td>2.0</td>\n",
1413 |        "      <td>0</td>\n",
1414 |        "      <td>2</td>\n",
1415 |        "      <td>0</td>\n",
1416 |        "      <td>0</td>\n",
1417 |        "      <td>...</td>\n",
1418 |        "      <td>0</td>\n",
1419 |        "      <td>0</td>\n",
1420 |        "      <td>1</td>\n",
1421 |        "      <td>0</td>\n",
1422 |        "      <td>0</td>\n",
1423 |        "      <td>1</td>\n",
1424 |        "      <td>0</td>\n",
1425 |        "      <td>0</td>\n",
1426 |        "      <td>0</td>\n",
1427 |        "      <td>0</td>\n",
1428 |        "    </tr>\n",
1429 |        "    <tr>\n",
1430 |        "      <th>7404</th>\n",
1431 |        "      <td>2007</td>\n",
1432 |        "      <td>3495</td>\n",
1433 |        "      <td>130000</td>\n",
1434 |        "      <td>200</td>\n",
1435 |        "      <td>44.8</td>\n",
1436 |        "      <td>2.0</td>\n",
1437 |        "      <td>1</td>\n",
1438 |        "      <td>2</td>\n",
1439 |        "      <td>4</td>\n",
1440 |        "      <td>0</td>\n",
1441 |        "      <td>...</td>\n",
1442 |        "      <td>0</td>\n",
1443 |        "      <td>0</td>\n",
1444 |        "      <td>0</td>\n",
1445 |        "      <td>1</td>\n",
1446 |        "      <td>0</td>\n",
1447 |        "      <td>0</td>\n",
1448 |        "      <td>0</td>\n",
1449 |        "      <td>0</td>\n",
1450 |        "      <td>0</td>\n",
1451 |        "      <td>1</td>\n",
1452 |        "    </tr>\n",
1453 |        "    <tr>\n",
1454 |        "      <th>9804</th>\n",
1455 |        "      <td>2012</td>\n",
1456 |        "      <td>5295</td>\n",
1457 |        "      <td>95360</td>\n",
1458 |        "      <td>30</td>\n",
1459 |        "      <td>62.8</td>\n",
1460 |        "      <td>2.0</td>\n",
1461 |        "      <td>1</td>\n",
1462 |        "      <td>0</td>\n",
1463 |        "      <td>0</td>\n",
1464 |        "      <td>1</td>\n",
1465 |        "      <td>...</td>\n",
1466 |        "      <td>0</td>\n",
1467 |        "      <td>0</td>\n",
1468 |        "      <td>0</td>\n",
1469 |        "      <td>1</td>\n",
1470 |        "      <td>0</td>\n",
1471 |        "      <td>1</td>\n",
1472 |        "      <td>0</td>\n",
1473 |        "      <td>0</td>\n",
1474 |        "      <td>0</td>\n",
1475 |        "      <td>0</td>\n",
1476 |        "    </tr>\n",
1477 |        "    <tr>\n",
1478 |        "      <th>6878</th>\n",
1479 |        "      <td>2017</td>\n",
1480 |        "      <td>19460</td>\n",
1481 |        "      <td>11076</td>\n",
1482 |        "      <td>145</td>\n",
1483 |        "      <td>56.5</td>\n",
1484 |        "      <td>3.0</td>\n",
1485 |        "      <td>0</td>\n",
1486 |        "      <td>2</td>\n",
1487 |        "      <td>0</td>\n",
1488 |        "      <td>0</td>\n",
1489 |        "      <td>...</td>\n",
1490 |        "      <td>0</td>\n",
1491 |        "      <td>0</td>\n",
1492 |        "      <td>1</td>\n",
1493 |        "      <td>0</td>\n",
1494 |        "      <td>0</td>\n",
1495 |        "      <td>1</td>\n",
1496 |        "      <td>0</td>\n",
1497 |        "      <td>0</td>\n",
1498 |        "      <td>0</td>\n",
1499 |        "      <td>0</td>\n",
1500 |        "    </tr>\n",
1501 |        "    <tr>\n",
1502 |        "      <th>10395</th>\n",
1503 |        "      <td>2017</td>\n",
1504 |        "      <td>15600</td>\n",
1505 |        "      <td>29019</td>\n",
1506 |        "      <td>145</td>\n",
1507 |        "      <td>65.7</td>\n",
1508 |        "      <td>2.0</td>\n",
1509 |        "      <td>0</td>\n",
1510 |        "      <td>0</td>\n",
1511 |        "      <td>0</td>\n",
1512 |        "      <td>1</td>\n",
1513 |        "      <td>...</td>\n",
1514 |        "      <td>0</td>\n",
1515 |        "      <td>0</td>\n",
1516 |        "      <td>1</td>\n",
1517 |        "      <td>0</td>\n",
1518 |        "      <td>0</td>\n",
1519 |        "      <td>1</td>\n",
1520 |        "      <td>0</td>\n",
1521 |        "      <td>0</td>\n",
1522 |        "      <td>0</td>\n",
1523 |        "      <td>0</td>\n",
1524 |        "    </tr>\n",
1525 |        "  </tbody>\n",
1526 |        "</table>\n",
1527 |        "<p>10 rows × 41 columns</p>\n",
1528 |        "</div>"
1529 |       ],
1530 |       "text/plain": [
1531 |        "       year  price  mileage  tax   mpg  engineSize  transmission_  model_  \\\n",
1532 |        "646    2019  29998    10171  145  41.5         2.0              2       2   \n",
1533 |        "9818   2016  17999    27193  160  50.4         3.0              0       4   \n",
1534 |        "8723   2017  13000    24104  125  53.3         1.5              1       0   \n",
1535 |        "3323   2019  50990     4729  145  37.7         3.0              2      17   \n",
1536 |        "2641   2019  30326     3392  145  41.5         2.0              2       2   \n",
1537 |        "2205   2015  10999    90018   20  70.6         2.0              0       2   \n",
1538 |        "7404   2007   3495   130000  200  44.8         2.0              1       2   \n",
1539 |        "9804   2012   5295    95360   30  62.8         2.0              1       0   \n",
1540 |        "6878   2017  19460    11076  145  56.5         3.0              0       2   \n",
1541 |        "10395  2017  15600    29019  145  65.7         2.0              0       0   \n",
1542 |        "\n",
1543 |        "       fuelType_  model_ 1 Series  ...  model_ i3  model_ i8  \\\n",
1544 |        "646            4                0  ...          0          0   \n",
1545 |        "9818           0                0  ...          0          0   \n",
1546 |        "8723           4                1  ...          0          0   \n",
1547 |        "3323           0                0  ...          0          0   \n",
1548 |        "2641           4                0  ...          0          0   \n",
1549 |        "2205           0                0  ...          0          0   \n",
1550 |        "7404           4                0  ...          0          0   \n",
1551 |        "9804           0                1  ...          0          0   \n",
1552 |        "6878           0                0  ...          0          0   \n",
1553 |        "10395          0                1  ...          0          0   \n",
1554 |        "\n",
1555 |        "       transmission_Automatic  transmission_Manual  transmission_Semi-Auto  \\\n",
1556 |        "646                         0                    0                       1   \n",
1557 |        "9818                        1                    0                       0   \n",
1558 |        "8723                        0                    1                       0   \n",
1559 |        "3323                        0                    0                       1   \n",
1560 |        "2641                        0                    0                       1   \n",
1561 |        "2205                        1                    0                       0   \n",
1562 |        "7404                        0                    1                       0   \n",
1563 |        "9804                        0                    1                       0   \n",
1564 |        "6878                        1                    0                       0   \n",
1565 |        "10395                       1                    0                       0   \n",
1566 |        "\n",
1567 |        "       fuelType_Diesel  fuelType_Electric  fuelType_Hybrid  fuelType_Other  \\\n",
1568 |        "646                  0                  0                0               0   \n",
1569 |        "9818                 1                  0                0               0   \n",
1570 |        "8723                 0                  0                0               0   \n",
1571 |        "3323                 1                  0                0               0   \n",
1572 |        "2641                 0                  0                0               0   \n",
1573 |        "2205                 1                  0                0               0   \n",
1574 |        "7404                 0                  0                0               0   \n",
1575 |        "9804                 1                  0                0               0   \n",
1576 |        "6878                 1                  0                0               0   \n",
1577 |        "10395                1                  0                0               0   \n",
1578 |        "\n",
1579 |        "       fuelType_Petrol  \n",
1580 |        "646                  1  \n",
1581 |        "9818                 0  \n",
1582 |        "8723                 1  \n",
1583 |        "3323                 0  \n",
1584 |        "2641                 1  \n",
1585 |        "2205                 0  \n",
1586 |        "7404                 1  \n",
1587 |        "9804                 0  \n",
1588 |        "6878                 0  \n",
1589 |        "10395                0  \n",
1590 |        "\n",
1591 |        "[10 rows x 41 columns]"
1592 |       ]
1593 |      },
1594 |      "metadata": {},
1595 |      "output_type": "display_data"
1596 |     },
1597 |     {
1598 |      "name": "stdout",
1599 |      "output_type": "stream",
1600 |      "text": [
1601 |       "(10781, 41)\n"
1602 |      ]
1603 |     }
1604 |    ],
1605 |    "source": [
1606 |     "print(df.shape)\n",
1607 |     "df = pd.get_dummies(df)\n",
1608 |     "# df = pd.get_dummies(df, drop_first=True)\n",
1609 |     "\n",
1610 |     "display(df.sample(10))\n",
1611 |     "print(df.shape)"
1612 |    ]
1613 |   },
1614 |   {
1615 |    "cell_type": "code",
1616 |    "execution_count": 62,
1617 |    "metadata": {},
1618 |    "outputs": [],
1619 |    "source": [
1620 |     "# Linear Regression\n",
1621 |     "# Decision Tree\n",
1622 |     "# Random Forest\n",
1623 |     "# XGB Xtreme Gradient Boosting\n",
1624 |     "# y = m1*x1 + m2*x2 + c\n",
1625 |     "\n",
1626 |     "# 1000\n",
1627 |     "# 80% training data = 800\n",
1628 |     "# 20% test data = 200 # seprate actual price\n",
1629 |     "\n",
1630 |     "# price = c1*model_1series + c2*model_2series + c3*year + c # ML training\n",
1631 |     "# predicted price = c1*model_1series + c2*model_2series + c3*year + c # ML testing\n",
1632 |     "\n",
1633 |     "# error = compare(actual price, predicted price)\n",
1634 |     "\n",
1635 |     "# 10%\n",
1636 |     "\n",
1637 |     "# 14%"
1638 |    ]
1639 |   },
1640 |   {
1641 |    "cell_type": "markdown",
1642 |    "metadata": {},
1643 |    "source": [
1644 |     "# Summary of Data Preprocessing for ML with Python\n",
1645 |     "What you have learned from this module:\n",
1646 |     "\n",
1647 |     "- Importing Data (csv, xlsx, txt etc.) with Pandas\n",
1648 |     "- creating a new DataFrame\n",
1649 |     "- column splitting\n",
1650 |     "- creating a new column in a dataframe\n",
1651 |     "- replace/removing a value from a pandas column\n",
1652 |     "- removing a column from the dataframe\n",
1653 |     "- renaming column names\n",
1654 |     "- extracting new information from a column\n",
1655 |     "- creating a column based on a condition or function\n",
1656 |     "- Removing a string from a column\n",
1657 |     "- Checking the unique values for each column\n",
1658 |     "- performing calculation in dataframe columns\n",
1659 |     "- dataframe sorting\n",
1660 |     "- dataframe slicing\n",
1661 |     "- data cleaning\n",
1662 |     "- data visualization of missing values\n",
1663 |     "- string to datetime conversion\n",
1664 |     "- removing missing values\n",
1665 |     "- replacing missing values by: 1. mean, 2. median, 3. constant, 4. interpolation, 5. forward imputation, 6. backward imputation\n",
1666 |     "- inner join, outer join, left join, right join\n",
1667 |     "- Data filtering\n",
1668 |     "- Data Aggregation/grouping \n",
1669 |     "- Pivot table\n",
1670 |     "- Data Visualization: Barplot\n",
1671 |     "- Dealing with categorical variables: 1. Label encoding, 2. One-hot encoding"
1672 |    ]
1673 |   },
1674 |   {
1675 |    "cell_type": "code",
1676 |    "execution_count": null,
1677 |    "metadata": {},
1678 |    "outputs": [],
1679 |    "source": []
1680 |   }
1681 |  ],
1682 |  "metadata": {
1683 |   "kernelspec": {
1684 |    "display_name": "Python 3 (ipykernel)",
1685 |    "language": "python",
1686 |    "name": "python3"
1687 |   },
1688 |   "language_info": {
1689 |    "codemirror_mode": {
1690 |     "name": "ipython",
1691 |     "version": 3
1692 |    },
1693 |    "file_extension": ".py",
1694 |    "mimetype": "text/x-python",
1695 |    "name": "python",
1696 |    "nbconvert_exporter": "python",
1697 |    "pygments_lexer": "ipython3",
1698 |    "version": "3.8.5"
1699 |   }
1700 |  },
1701 |  "nbformat": 4,
1702 |  "nbformat_minor": 4
1703 | }
1704 | 


--------------------------------------------------------------------------------
/ODI_cricket.xlsx:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/SKawsar/Data-Preprocessing-for-ML/7918aaa43c3594ed9781d72fd438f60da9247eae/ODI_cricket.xlsx


--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
 1 | I taught Data Preprocessing with Python to 200+ students through an online platform. These are the course materials I built for my students (Mostly graduate-level students from the Non-CS background).
 2 | 
 3 | Video Lectures on YouTube: https://lnkd.in/gK5epasB
 4 | 
 5 | Medium blogs for pandas: https://kawsar34.medium.com/list/learn-pandas-from-leetcode-a06903853aed
 6 | 
 7 | # Data Preprocessing with Python
 8 | 
 9 | #### Lecture 01: Importing Data with Pandas
10 | - Challenges of reading a CSV file
11 | - Understanding the data
12 | - Finding Data Statistics, data types and missing value information
13 | 
14 | #### Lecture 02: Data Preprocessing with Pandas (Part 1)
15 | - Challenges of reading a CSV or Excel file
16 | - Choose columns by name before reading a csv file
17 | - Choose columns by number before reading a csv file
18 | - Reading only the first n number of rows
19 | 
20 | #### Lecture 03: Data Preproccessing with Pandas (Part 2)
21 | - How to extract new information from a column?
22 | - How to create a column based on a condition or function?
23 | - Removing a string from a column
24 | - Checking the unique values for each column
25 | - performing calculation in dataframe columns
26 | - dataframe sorting
27 | 
28 | #### Lecture 04: HW review
29 | 
30 | #### Lecture 05: Handling Missing values
31 | - performing data cleaning
32 | - data visualization of missing values
33 | - string to datetime conversion
34 | - removing missing values
35 | - replacing missing values by: 1. mean, 2. median, 3. constant, 4. interpolation, 5. forward imputation, 6. backward imputation
36 | 
37 | #### Lecture 06: Data Joining using Pandas
38 | - inner join, outer join, left join, right join
39 | 
40 | #### Lecture 07: Data Aggregation/grouping and Pivot table using Pandas
41 | - Data filtering
42 | - Data preprocessing
43 | - Data Aggregation/grouping 
44 | - Pivot table
45 | - Data Visualization: Barplot
46 | 
47 | #### Lecture 08: Data Correlation and Categorical Variable Encoding
48 | - Dealing with categorical variables
49 | - Label encoding
50 | - One-hot encoding
51 | 


--------------------------------------------------------------------------------
/batsman_most_runs_ODI.csv:
--------------------------------------------------------------------------------
 1 | Player,Span,Mat,Inns,NO,Runs,HS,Ave,BF,SR,100,50,0,4s,6s
 2 | SR Tendulkar (INDIA),1989-2012,463,452,41,18426,200*,44.83,21368,86.23,49,96,20,2016,195
 3 | KC Sangakkara (Asia/ICC/SL),2000-2015,404,380,41,14234,169,41.98,18048,78.86,25,93,15,1385,88
 4 | RT Ponting (AUS/ICC),1995-2012,375,365,39,13704,164,42.03,17046,80.39,30,82,20,1231,162
 5 | ST Jayasuriya (Asia/SL),1989-2011,445,433,18,13430,189,32.36,14725,91.2,28,68,34,1500,270
 6 | DPMD Jayawardene (Asia/SL),1998-2015,448,418,39,12650,144,33.37,16020,78.96,19,77,28,1119,76
 7 | V Kohli (INDIA),2008-2022,260,251,39,12311,183,58.07,13249,92.92,43,64,15,1153,125
 8 | Inzamam-ul-Haq (Asia/PAK),1991-2007,378,350,53,11739,137*,39.52,15812,74.24,10,83,20,971,144
 9 | JH Kallis (Afr/ICC/SA),1996-2014,328,314,53,11579,139,44.36,15885,72.89,17,86,17,911,137
10 | SC Ganguly (Asia/INDIA),1992-2007,311,300,23,11363,183,41.02,15416,73.7,22,72,16,1122,190
11 | R Dravid (Asia/ICC/INDIA),1996-2011,344,318,40,10889,153,39.16,15285,71.23,12,83,13,950,42
12 | MS Dhoni (Asia/INDIA),2004-2019,350,297,84,10773,183*,50.57,12303,87.56,10,73,10,826,229
13 | CH Gayle (ICC/WI),1999-2019,301,294,17,10480,215,37.83,12019,87.19,25,54,25,1128,331
14 | BC Lara (ICC/WI),1990-2007,299,289,32,10405,169,40.48,13086,79.51,19,63,16,1042,133
15 | TM Dilshan (SL),1999-2016,330,303,41,10290,161*,39.27,11933,86.23,22,47,11,1111,55
16 | Mohammad Yousuf (Asia/PAK),1998-2010,288,273,40,9720,141*,41.71,12942,75.1,15,64,15,785,90
17 | AC Gilchrist (AUS/ICC),1996-2008,287,279,11,9619,172,35.89,9922,96.94,16,55,19,1162,149
18 | AB de Villiers (Afr/SA),2005-2018,228,218,39,9577,176,53.5,9473,101.09,25,53,7,840,204
19 | M Azharuddin (INDIA),1985-2000,334,308,54,9378,153*,36.92,12669,74.02,7,58,9,622+,77+
20 | PA de Silva (SL),1984-2003,308,296,30,9284,145,34.9,11443,81.13,11,64,17,712+,102+
21 | RG Sharma (INDIA),2007-2022,230,223,32,9283,264,48.6,10428,89.01,29,44,13,845,245
22 | Saeed Anwar (PAK),1989-2003,247,244,19,8824,194,39.21,10938,80.67,20,43,15,938,97
23 | S Chanderpaul (WI),1994-2011,268,251,40,8778,150,41.6,12408,70.74,11,59,6,722,85
24 | Yuvraj Singh (Asia/INDIA),2000-2017,304,278,40,8701,150,36.55,9924,87.67,14,52,18,908,155
25 | DL Haynes (WI),1978-1994,238,237,28,8648,152*,41.37,13707,63.09,17,57,13,768+,53+
26 | LRPL Taylor (NZ),2006-2021,233,217,39,8581,181*,48.2,10287,83.41,21,51,9,712,146
27 | MS Atapattu (SL),1990-2007,268,259,32,8529,132*,37.57,12594,67.72,11,59,13,734,15
28 | ME Waugh (AUS),1988-2002,244,236,20,8500,173,39.35,11053,76.9,18,50,16,651,57
29 | V Sehwag (Asia/ICC/INDIA),1999-2013,251,245,9,8273,219,35.05,7929,104.33,15,38,14,1132,136
30 | HM Amla (SA),2008-2019,181,178,14,8113,159,49.46,9178,88.39,27,39,4,822,53
31 | HH Gibbs (SA),1996-2010,248,240,16,8094,175,36.13,9721,83.26,21,37,22,930,128
32 | Shahid Afridi (Asia/ICC/PAK),1996-2015,398,369,27,8064,124,23.57,6892,117,6,39,30,730,351
33 | SP Fleming (ICC/NZ),1994-2007,280,269,21,8037,134*,32.4,11242,71.49,8,49,17,823,63
34 | MJ Clarke (AUS),2003-2015,245,223,44,7981,130,44.58,10104,78.98,8,58,10,665,53
35 | Tamim Iqbal (BAN),2007-2022,225,223,10,7826,158,36.74,9963,78.55,14,52,19,853,99
36 | EJG Morgan (ENG/IRE),2006-2021,246,228,34,7701,148,39.69,8439,91.25,14,47,16,654,220
37 | SR Waugh (AUS),1986-2002,325,288,58,7569,120*,32.9,9971,75.91,3,45,15,530,68
38 | Shoaib Malik (PAK),1999-2019,287,258,40,7534,143,34.55,9199,81.9,9,44,15,603,113
39 | A Ranatunga (SL),1982-1999,269,255,47,7456,131*,35.84,9571,77.9,4,49,18,523+,64+
40 | Javed Miandad (PAK),1975-1996,233,218,41,7381,119*,41.7,11014,67.01,8,50,8,445+,44+
41 | Younis Khan (PAK),2000-2015,265,255,23,7249,144,31.24,9628,75.29,7,48,22,578,56
42 | Saleem Malik (PAK),1982-1999,283,256,38,7170,102,32.88,9383,76.41,5,47,19,514+,34+
43 | NJ Astle (NZ),1995-2007,223,217,14,7090,145*,34.92,9760,72.64,16,41,19,720,86
44 | GC Smith (Afr/SA),2002-2013,197,194,10,6989,141,37.98,8648,80.81,10,47,8,788,44
45 | WU Tharanga (Asia/SL),2005-2019,235,223,17,6951,174*,33.74,9155,75.92,15,37,17,798,52
46 | MJ Guptill (NZ),2009-2021,186,183,19,6927,237*,42.23,7896,87.72,16,37,15,702,181
47 | MG Bevan (AUS),1994-2004,232,196,67,6912,108*,53.58,9320,74.16,6,46,5,450,21
48 | G Kirsten (SA),1993-2003,185,185,19,6798,188*,40.95,9436,72.04,13,45,11,659,20
49 | A Flower (ZIM),1992-2003,213,208,16,6786,145,35.34,9097,74.59,4,55,13,532+,26+
50 | Shakib Al Hasan (BAN),2006-2022,221,209,30,6755,134*,37.73,8212,82.25,9,50,12,611,46
51 | IVA Richards (WI),1975-1991,187,167,24,6721,189*,47,7451,90.2,11,45,7,600+,126+
52 | Mushfiqur Rahim (BAN),2006-2022,233,218,36,6697,144,36.79,8481,78.96,8,41,10,527,85
53 | BRM Taylor (ZIM),2004-2021,205,203,15,6684,145*,35.55,8721,76.64,11,39,15,599,106
54 | Mohammad Hafeez (PAK),2003-2019,218,216,15,6614,140*,32.9,8633,76.61,11,38,19,664,110
55 | GW Flower (ZIM),1992-2010,221,214,18,6571,142*,33.52,9723,67.58,6,40,18,557+,37+
56 | Ijaz Ahmed (PAK),1986-2000,250,232,29,6564,139*,32.33,8174,80.3,10,37,14,531+,87+
57 | AR Border (AUS),1979-1994,273,252,39,6524,127*,30.62,9134,71.42,3,39,11,500,43
58 | S Dhawan (INDIA),2010-2022,149,146,8,6284,143,45.53,6730,93.37,17,35,5,783,74
59 | RB Richardson (WI),1983-1996,224,217,30,6248,122,33.41,9801,63.74,5,44,8,541+,54+
60 | KS Williamson (NZ),2010-2020,151,144,14,6173,148,47.48,7551,81.75,13,39,5,563,49
61 | ML Hayden (AUS/ICC),1993-2008,161,155,15,6133,181*,43.8,7767,78.96,10,36,9,636,87
62 | JE Root (ENG),2013-2021,152,142,23,6109,133*,51.33,7034,86.84,16,35,5,491,44
63 | BB McCullum (NZ),2002-2016,260,228,28,6083,166,30.41,6312,96.37,5,32,20,577,200
64 | DM Jones (AUS),1984-1994,164,161,25,6068,145,44.61,8362,72.56,7,46,6,380+,64+
65 | DC Boon (AUS),1984-1995,181,177,16,5964,122,37.04,9157,65.13,5,37,6,494,16
66 | JN Rhodes (SA),1992-2003,245,220,51,5935,121,35.11,7336,80.9,2,33,12,392,47
67 | Ramiz Raja (PAK),1985-1997,198,197,15,5841,119*,32.09,9226,63.31,9,31,15,469+,14+
68 | AD Mathews (SL),2008-2021,218,188,48,5835,139*,41.67,7004,83.3,3,40,15,461,89
69 | RR Sarwan (WI),2000-2013,181,169,33,5804,120*,42.67,7663,75.74,5,38,8,480,58
70 | CL Hooper (WI),1987-2003,227,206,43,5761,113*,35.34,7517,76.63,7,29,7,409,65
71 | SR Watson (AUS),2002-2015,190,169,27,5757,185*,40.54,6365,90.44,9,33,12,570,131
72 | Q de Kock (SA),2013-2022,129,129,6,5658,178,46,5895,95.97,17,28,4,646,88
73 | H Masakadza (ZIM),2001-2019,209,208,4,5658,178*,27.73,7728,73.21,5,34,15,585,86
74 | SK Raina (INDIA),2005-2018,226,194,35,5615,116*,35.31,6005,93.5,5,36,14,476,120
75 | MN Samuels (WI),2000-2018,207,196,26,5606,133*,32.97,7463,75.11,10,30,11,526,118
76 | WJ Cronje (SA),1992-2000,188,175,31,5565,112,38.64,7277,76.47,2,39,8,366,94
77 | F du Plessis (SA),2011-2019,143,136,20,5507,185,47.47,6215,88.6,12,35,3,495,66
78 | DA Warner (AUS),2009-2020,128,126,6,5455,179,45.45,5710,95.53,18,23,2,571,85
79 | MEK Hussey (AUS),2004-2012,185,157,44,5442,109*,48.15,6243,87.16,3,39,3,383,80
80 | IR Bell (ENG),2004-2015,161,157,14,5416,141,37.87,7019,77.16,4,35,6,525,32
81 | A Jadeja (INDIA),1992-2000,196,179,36,5359,119,37.47,7678,69.79,6,30,10,366,85
82 | DR Martyn (AUS),1992-2006,208,182,51,5346,144*,40.8,6877,77.73,5,37,10,441,22
83 | G Gambhir (INDIA),2003-2013,147,143,11,5238,150*,39.68,6144,85.25,11,34,11,561,17
84 | AJ Finch (AUS),2013-2020,132,128,3,5232,153*,41.85,5917,88.42,17,29,11,518,126
85 | ADR Campbell (ZIM),1992-2003,188,184,14,5185,131*,30.5,7834,66.18,7,30,11,427+,44+
86 | RS Mahanama (SL),1986-1999,213,198,23,5162,119*,29.49,8521,60.57,4,35,15,331+,17+
87 | CG Greenidge (WI),1975-1991,128,127,13,5134,133*,45.03,7908,64.92,11,31,3,470+,81+
88 | Misbah-ul-Haq (PAK),2002-2015,162,149,31,5122,96*,43.4,6945,73.75,0,42,6,342,83
89 | JP Duminy (SA),2004-2019,199,179,40,5117,150*,36.81,6052,84.55,4,27,7,347,75
90 | PD Collingwood (ENG),2001-2011,197,181,37,5092,120*,35.36,6614,76.98,5,26,7,365,74
91 | A Symonds (AUS),1998-2009,198,161,33,5088,156,39.75,5504,92.44,6,30,15,449,103
92 | Abdul Razzaq (Asia/PAK),1996-2011,265,228,57,5080,112,29.7,6252,81.25,3,23,14,382,124
93 | PR Stirling (IRE),2008-2022,136,133,3,5047,177,38.82,5840,86.42,12,26,10,541,114


--------------------------------------------------------------------------------
/bowler_most_wickets_ODI.csv:
--------------------------------------------------------------------------------
 1 | Player,Span,Mat,Inns,Balls,Runs,Wkts,Ave,Econ,SR,4,5
 2 | M Muralitharan (Asia/ICC/SL),1993-2011,350,341,18811,12326,534,23.08,3.93,35.2,15,10
 3 | Wasim Akram (PAK),1984-2003,356,351,18186,11812,502,23.52,3.89,36.2,17,6
 4 | Waqar Younis (PAK),1989-2003,262,258,12698,9919,416,23.84,4.68,30.5,14,13
 5 | WPUJC Vaas (Asia/SL),1994-2008,322,320,15775,11014,400,27.53,4.18,39.4,9,4
 6 | Shahid Afridi (Asia/ICC/PAK),1996-2015,398,372,17670,13632,395,34.51,4.62,44.7,4,9
 7 | SM Pollock (Afr/ICC/SA),1996-2008,303,297,15712,9631,393,24.5,3.67,39.9,12,5
 8 | GD McGrath (AUS/ICC),1993-2007,250,248,12970,8391,381,22.02,3.88,34,9,7
 9 | B Lee (AUS),2000-2012,221,217,11185,8877,380,23.36,4.76,29.4,14,9
10 | SL Malinga (SL),2004-2019,226,220,10936,9760,338,28.87,5.35,32.3,11,8
11 | A Kumble (Asia/INDIA),1990-2007,271,265,14496,10412,337,30.89,4.3,43,8,2
12 | ST Jayasuriya (Asia/SL),1989-2011,445,368,14874,11871,323,36.75,4.78,46,8,4
13 | J Srinath (INDIA),1991-2003,229,227,11935,8847,315,28.08,4.44,37.8,7,3
14 | DL Vettori (ICC/NZ),1997-2015,295,277,14060,9674,305,31.71,4.12,46,8,2
15 | SK Warne (AUS/ICC),1993-2005,194,191,10642,7541,293,25.73,4.25,36.3,12,1
16 | Saqlain Mushtaq (PAK),1995-2003,169,165,8770,6275,288,21.78,4.29,30.4,11,6
17 | AB Agarkar (INDIA),1998-2007,191,188,9484,8021,288,27.85,5.07,32.9,10,2
18 | Shakib Al Hasan (BAN),2006-2022,221,218,11351,8401,285,29.47,4.44,39.8,9,3
19 | Z Khan (Asia/INDIA),2000-2012,200,197,10097,8301,282,29.43,4.93,35.8,7,1
20 | JH Kallis (Afr/ICC/SA),1996-2014,328,283,10750,8680,273,31.79,4.84,39.3,2,2
21 | AA Donald (SA),1991-2003,164,162,8561,5926,272,21.78,4.15,31.4,11,2
22 | Mashrafe Mortaza (Asia/BAN),2001-2020,220,220,10922,8893,270,32.93,4.88,40.4,7,1
23 | JM Anderson (ENG),2002-2015,194,191,9584,7861,269,29.22,4.92,35.6,11,2
24 | Abdul Razzaq (Asia/PAK),1996-2011,265,254,10941,8564,269,31.83,4.69,40.6,8,3
25 | Harbhajan Singh (Asia/INDIA),1998-2015,236,227,12479,8973,269,33.35,4.31,46.3,2,3
26 | M Ntini (ICC/SA),1998-2009,173,171,8687,6559,266,24.65,4.53,32.6,8,4
27 | N Kapil Dev (INDIA),1978-1994,225,221,11202,6945,253,27.45,3.71,44.2,3,1
28 | Shoaib Akhtar (Asia/ICC/PAK),1998-2011,163,162,7764,6169,247,24.97,4.76,31.4,6,4
29 | KD Mills (NZ),2001-2015,170,169,8230,6485,240,27.02,4.72,34.2,8,1
30 | MG Johnson (AUS),2005-2015,153,150,7489,6038,239,25.26,4.83,31.3,9,3
31 | HH Streak (Afr/ZIM),1993-2005,189,185,9468,7129,239,29.82,4.51,39.6,7,1
32 | D Gough (ENG/ICC),1994-2006,159,156,8470,6209,235,26.42,4.39,36,10,2
33 | CA Walsh (WI),1985-2000,205,204,10822,6918,227,30.47,3.83,47.6,6,1
34 | CEL Ambrose (WI),1988-2000,176,175,9353,5429,225,24.12,3.48,41.5,6,4
35 | Abdur Razzak (BAN),2004-2014,153,152,7965,6065,207,29.29,4.56,38.4,5,4
36 | CJ McDermott (AUS),1985-1996,138,138,7461,5018,203,24.71,4.03,36.7,4,1
37 | CZ Harris (NZ),1990-2004,250,232,10667,7613,203,37.5,4.28,52.5,2,1
38 | CL Cairns (ICC/NZ),1991-2006,215,186,8168,6594,201,32.8,4.84,40.6,3,1
39 | DJ Bravo (WI),2004-2014,164,150,6511,5874,199,29.51,5.41,32.7,6,1
40 | KMDN Kulasekara (SL),2003-2017,184,181,8263,6751,199,33.92,4.9,41.5,4,1
41 | DW Steyn (Afr/SA),2005-2019,125,124,6256,5087,196,25.95,4.87,31.9,4,3
42 | BKV Prasad (INDIA),1994-2001,161,160,8129,6332,196,32.3,4.67,41.4,3,1
43 | MA Starc (AUS),2010-2021,99,99,5099,4379,195,22.45,5.15,26.1,11,8
44 | SR Waugh (AUS),1986-2002,325,207,8883,6761,195,34.67,4.56,45.5,3,0
45 | CL Hooper (WI),1987-2003,227,203,9573,6958,193,36.05,4.36,49.6,3,0
46 | L Klusener (SA),1996-2004,171,164,7336,5751,192,29.95,4.7,38.2,1,6
47 | TG Southee (NZ),2008-2020,143,141,7195,6558,190,34.51,5.46,37.8,4,3
48 | M Morkel (Afr/SA),2007-2018,117,114,5760,4761,188,25.32,4.95,30.6,7,2
49 | RA Jadeja (INDIA),2009-2020,168,164,8557,7024,188,37.36,4.92,45.5,7,1
50 | CRD Fernando (Asia/SL),2001-2012,147,141,6507,5648,187,30.2,5.2,34.7,3,1
51 | Saeed Ajmal (PAK),2008-2015,113,112,6000,4182,184,22.72,4.18,32.6,6,2
52 | Imran Khan (PAK),1974-1992,175,153,7461,4844,182,26.61,3.89,40.9,3,1
53 | Aaqib Javed (PAK),1988-1998,163,159,8012,5721,182,31.43,4.28,44,2,4
54 | Umar Gul (PAK),2003-2016,130,128,6064,5253,179,29.34,5.19,33.8,4,2
55 | SCJ Broad (ENG),2006-2016,121,121,6109,5364,178,30.13,5.26,34.3,9,1
56 | NLTC Perera (SL),2009-2021,166,157,5900,5740,175,32.8,5.83,33.7,5,4
57 | NW Bracken (AUS),2001-2009,116,116,5759,4240,174,24.36,4.41,33,5,2
58 | Imran Tahir (SA),2011-2019,107,104,5541,4297,173,24.83,4.65,32,7,3
59 | JDP Oram (NZ),2001-2012,160,154,6911,5047,173,29.17,4.38,39.9,3,2
60 | IK Pathan (INDIA),2004-2012,120,118,5855,5142,173,29.72,5.26,33.8,5,2
61 | A Flintoff (ENG/ICC),1999-2009,141,119,5624,4121,169,24.38,4.39,33.2,6,2
62 | TA Boult (NZ),2012-2021,93,93,5117,4261,169,25.21,4.99,30.2,8,5
63 | SR Watson (AUS),2002-2015,190,163,6466,5342,168,31.79,4.95,38.4,3,0
64 | CH Gayle (ICC/WI),1999-2019,301,199,7424,5926,167,35.48,4.78,44.4,3,1
65 | Mushtaq Ahmed (PAK),1989-2003,144,142,7543,5361,161,33.29,4.26,46.8,3,1
66 | AU Rashid (ENG),2009-2021,112,106,5573,5251,159,33.02,5.65,35,7,2
67 | RJ Hadlee (NZ),1973-1990,115,112,6182,3407,158,21.56,3.3,39.1,1,5
68 | Shoaib Malik (PAK),1999-2019,287,217,7958,6192,158,39.18,4.66,50.3,1,0
69 | MD Marshall (WI),1980-1992,136,134,7175,4233,157,26.96,3.53,45.7,6,0
70 | M Prabhakar (INDIA),1984-1996,130,127,6360,4534,157,28.87,4.27,40.5,4,2
71 | A Nehra (Asia/INDIA),2001-2011,120,120,5751,4981,157,31.72,5.19,36.6,5,2
72 | GB Hogg (AUS),1996-2008,123,113,5564,4188,156,26.84,4.51,35.6,3,2
73 | CR Woakes (ENG),2011-2021,106,102,5016,4567,155,29.46,5.46,32.3,10,3
74 | SR Tendulkar (INDIA),1989-2012,463,270,8054,6850,154,44.48,5.1,52.2,4,2
75 | BAW Mendis (SL),2008-2015,87,84,4154,3324,152,21.86,4.8,27.3,7,3
76 | Rashid Khan (AFG),2015-2022,80,76,4074,2821,151,18.68,4.15,26.9,5,4
77 | UDU Chandana (SL),1994-2007,147,136,6142,4818,151,31.9,4.7,40.6,4,1
78 | R Ashwin (INDIA),2010-2022,113,111,6141,5058,151,33.49,4.94,40.6,1,0
79 | 


--------------------------------------------------------------------------------
/gre.csv:
--------------------------------------------------------------------------------
 1 | Date,verbal_score,quant_score
 2 | 09/01/2021,0,1
 3 | 09/02/2021,1,2
 4 | 09/03/2021,2,3
 5 | 09/04/2021,3,
 6 | 09/05/2021,4,
 7 | 09/06/2021,,
 8 | 09/07/2021,,7
 9 | 09/08/2021,7,8
10 | 09/09/2021,8,9
11 | 09/10/2021,9,10


--------------------------------------------------------------------------------